Health Care Spending and the€¦ · share, private health insurance is not a single purchaser of...
Transcript of Health Care Spending and the€¦ · share, private health insurance is not a single purchaser of...
Health Care Spending
and the
Medicare Program
iii
The MedPAC Data Book provides information on national health care and Medicare spending as
well as Medicare beneficiary demographics, dual-eligible beneficiaries, quality of care in the
Medicare program, and Medicare beneficiary and other payer liability. It also examines provider
settings—such as hospitals and post-acute care—and presents data on Medicare spending,
beneficiaries’ access to care in the setting (measured by the number of beneficiaries using the
service, number of providers, volume of services, length of stay, or through direct surveys), and
the sector’s Medicare profit margins, if applicable. In addition, it covers the Medicare Advantage
program and prescription drug coverage for Medicare beneficiaries, including Part D.
MedPAC began producing its annual Data Book at the suggestion of congressional staff. Some
of the information it contains is derived from MedPAC’s March and June reports to the
Congress; other information is unique to the Data Book. The information is presented in tables
and figures with brief discussions.
We produce a limited number of printed copies of this report. It is, however, available through
the MedPAC website: www.medpac.gov.
Changes in aggregate spending for the fee-for-service sectors presented in this Data Book partly
reflect the shift in Medicare enrollment from the traditional fee-for-service program to Medicare
Advantage. Fee-for-service spending per capita may present a more complete picture of spending
changes.
v
............................................................................................................................
1-1 Medicare was the largest single purchaser of personal health care, 2018 ...........................................
1-2 Medicare spending is concentrated in certain services and has shifted over time .............................
1-3 Aggregate Medicare spending for FFS beneficiaries, by sector, 2010–2019 ...................................
1-4 Per capita Medicare spending for FFS beneficiaries, by sector, 2010–2019 ....................................
1-5 Medicare’s share of spending on personal health care varied by type of service, 2018 ....................
1-6 Health care spending has consumed an increasing share of the country’s GDP ...............................
1-7 Trustees project Medicare spending to continue to increase as a share of GDP ...............................
1-8 Per beneficiary spending growth slowed in recent years but is projected to accelerate ..................
1-9 Employer-sponsored insurance premiums have risen more than twice as fast as
Medicare FFS costs ..................................................................................................................
1-10 Trustees and CBO project Medicare spending to exceed $1 trillion by 2022 ................................
1-11 FFS program spending was highly concentrated in a small group of beneficiaries, 2017 ..............
1-12 Medicare HI Trust Fund is projected to be depleted in 2026 under Trustees’
intermediate assumptions ..........................................................................................................
1-13 General revenues have overtaken Medicare payroll taxes as the largest source
of Medicare funding .................................................................................................................
1-14 Medicare enrollment is rising while the number of workers per HI beneficiary is declining .........
1-15 Medicare HI and SMI benefits and cost sharing per FFS beneficiary, 2018 .................................
2-1 Aged beneficiaries accounted for the greatest share of the Medicare population and
program spending, 2017............................................................................................................
2-2 Beneficiaries younger than 65 accounted for a disproportionate share of Medicare
spending, 2017 .........................................................................................................................
2-3 Beneficiaries who reported being in poor health accounted for a disproportionate share of
Medicare spending, 2017 .....................................................................................................................
2-4 Enrollment in the Medicare program is projected to grow rapidly through 2030 ..........................
2-5 Characteristics of the Medicare population, 2017 .......................................................................
3-1 Sources of supplemental coverage among noninstitutionalized Medicare beneficiaries, 2017 .......
3-2 Sources of supplemental coverage among noninstitutionalized Medicare beneficiaries,
by beneficiaries’ characteristics, 2017........................................................................................
vi
3-3 Covered benefits and enrollment in standardized Medigap plans, 2018 .......................................
3-4 Total spending on health care services for noninstitutionalized FFS Medicare
beneficiaries, by source of payment, 2017..................................................................................
3-5 Per capita total spending on health care services among noninstitutionalized FFS
beneficiaries, by source of payment, 2017..................................................................................
4-1 Dual-eligible beneficiaries accounted for a disproportionate share of Medicare spending, 2017 ...
4-2 Dual-eligible beneficiaries were more likely than non-dual-eligible beneficiaries to be
under age 65 and disabled, 2017 ................................................................................................
4-3 Dual-eligible beneficiaries were more likely than non-dual-eligible beneficiaries to report
being in poor health, 2017 .........................................................................................................
4-4 Demographic differences between dual-eligible beneficiaries and non-dual-eligible
beneficiaries, 2017 ....................................................................................................................
4-5 Differences in Medicare spending and service use between dual-eligible beneficiaries
and non-dual-eligible beneficiaries, 2017 ...................................................................................
4-6 Both Medicare and total spending were concentrated among dual-eligible beneficiaries, 2017 .....
5-1 Most beneficiaries are in Medicare managed care plans or are assigned to accountable
care organizations, 2020 ........................................................................................................... 5-2 The number of beneficiaries assigned to MSSP ACOs grew rapidly through 2018 but
more moderately since then ....................................................................................................... 5-3 Bundled Payments for Care Improvement (BPCI) Advanced is Medicare’s largest
episode-based payment model, 2020 ......................................................................................... 5-4 Share of BPCI Advanced participants accepting financial responsibility for particular types of
clinical episodes, 2020 ..............................................................................................................
5-5 2,825 practices are testing the Comprehensive Primary Care Plus (CPC+) model, 2020 ...............
5-6 About 95 percent of the clinicians who qualified for a 5 percent A–APM bonus in 2020
were in ACO models ................................................................................................................
6-1 Urban IPPS hospitals accounted for under half of the 4,700 short-term acute care hospitals
but about 80 percent of all-payer and Medicare FFS inpatient stays in 2018 ................................
6-2 The number of general, short-term acute care hospitals that ceased inpatient services
substantially increased in 2019 ..................................................................................................
6-3 Aggregate occupancy rate at short-term acute care hospitals has increased slightly, but
remained much lower at rural hospitals, 2014–2018 ...................................................................
6-4 Divergent trends in all-payer hospital outpatient visits and inpatient stays per capita
narrowed in 2017 and 2018 .......................................................................................................
vii
6-5 Hospital employment growth from 2015 to 2019 was driven by an increase in health
care practitioners ......................................................................................................................
6-6 Short-term acute care hospitals’ all-payer financial performance remained strong, 2014–2018 .....
6-7 Urban IPPS hospitals continue to have a higher aggregate total all-payer margin than
rural IPPS or critical access hospitals, 2014–2018 ......................................................................
6-8 Major teaching hospitals continue to have a lower aggregate total all-payer margin
than nonteaching and other teaching hospitals, 2014–2018 .........................................................
6-9 Inpatient services are the largest component of Medicare FFS revenue for IPPS hospitals,
while outpatient services are the largest and fastest growing for CAHs, 2014–2018 .....................
6-10 Base PPS payments represented about 84 percent of IPPS hospitals’ overall Medicare
FFS revenue, 2018 ....................................................................................................................
6-11 Since implementation of uncompensated care payments in 2014, DSH payments have slowly
increased, while uncompensated care payments have varied .......................................................
6-12 Financial pressure led to lower hospital costs per discharge, 2015–2017 .....................................
6-13 Aggregate overall Medicare margin for short-term acute care hospitals increased slightly
from 2017 to 2018 ....................................................................................................................
6-14 Rural short-term acute care hospitals continue to have a higher aggregate overall
Medicare margin than urban hospitals .......................................................................................
6-15 Teaching and disproportionate share short-term acute care hospitals continue to have higher
aggregate Medicare margins than other hospitals .......................................................................
6-16 Decline in Medicare FFS inpatient stays per capita has slowed, with larger declines at
critical access and rural hospitals ...............................................................................................
6-17 Circulatory system diagnoses remain the most common reason for Medicare FFS beneficiaries'
inpatient stays at short-term acute care hospitals, but the share has decreased, 2010–2018 ...........
6-18 A growing share of Medicare FFS beneficiaries' inpatient stays at short-term acute care
hospitals are immediately followed by post-acute care or hospice, 2010–2018 ............................
6-19 Share of Medicare FFS beneficiaries' short stays at short-term acute care hospitals
decreased from 2010 to 2014 but then began to increase again....................................................
6-20 Number of Medicare FFS outpatient observation visits per capita decreased slightly in 2017
and 2018, and nearly half remained longer than 24 hours ...........................................................
6-21 Growth in Medicare FFS hospital outpatient services per capita has slowed, 2010–2018 .............
6-22 Medicare FFS payments to inpatient psychiatric facilities decreased in 2018 ...............................
6-23 A growing share of Medicare-certified inpatient psychiatric facilities are for profit, 2011–2018 ...
6-24 Almost three-quarters of Medicare FFS beneficiaries’ stays at IPFs were for psychosis, 2018 ......
6-25 The majority of Medicare FFS beneficiaries who received IPF services were under the
age of 65, 2018 .........................................................................................................................
7-1 Medicare spending per fee-for-service beneficiary on services in the fee schedule for
physicians and other health professionals, 2009–2019 ................................................................
7-2 Physician fee schedule–allowed charges by type of service, 2018 ...............................................
7-3 Total encounters per beneficiary increased and mix of clinicians furnishing them
changed from 2013 to 2018 .......................................................................................................
7-4 Medicare beneficiaries’ ability to get timely appointments with physicians was comparable
with privately insured individuals, 2016–2019 ...........................................................................
viii
7-5 Medicare and privately insured patients who were looking for a new physician reported more
difficulty finding one in primary care, 2016–2019 ......................................................................
7-6 Medicare beneficiaries’ access to physician care was comparable with privately insured
individuals, and minorities in both groups reported unwanted delays more frequently, 2019 ........
7-7 Minorities in Medicare and with private insurance were more likely to report problems
finding a new specialist, 2019 ...................................................................................................
7-8 Changes in physicians’ professional liability insurance premiums, 2012–2019 ............................
7-9 Spending on hospital outpatient services covered under the outpatient PPS, 2009–2019 ..............
7-10 Most hospitals provide outpatient services .................................................................................
7-11 Payments and volume of services under the Medicare hospital outpatient PPS, by type of
service, 2018 ............................................................................................................................
7-12 Hospital outpatient services with the highest Medicare expenditures, 2018 .................................
7-13 Separately payable drugs have increased as a share of total spending in the outpatient
prospective payment system, 2013–2018 ...................................................................................
7-14 Number of Medicare-certified ASCs increased by 10 percent, 2012–2018 ..................................
8-1 The number of post-acute care providers decreased slightly in 2019 ...........................................
8-2 Medicare fee-for-service spending for post-acute care expenditures was relatively stable
from 2010 to 2018 ....................................................................................................................
8-3 Freestanding SNFs and for-profit SNFs accounted for the majority of facilities, Medicare
stays, and Medicare spending in 2012 and 2018 .........................................................................
8-4 SNF admissions and stays continued to decline in 2018 ...........................................................
8-5 Freestanding SNF Medicare margins remained high in 2018 ....................................................
8-6 Cost and payment differences explain variation in Medicare margins for freestanding SNFs
in 2018 ...................................................................................................................................
8-7 Financial performance of relatively efficient SNFs in 2018 reflects a combination of lower
cost per day and higher payment per day .................................................................................
8-8 SNFs improved on some measures but not others from 2012 to 2018 ........................................
8-9 Trends in the provision of home health care .............................................................................
8-10 Most home health episodes are not preceded by hospitalization or PAC stay .............................
8-11 Medicare margins for freestanding home health agencies, 2017 and 2018 .................................
8-12 Home health agencies’ assessment-based performance measures increased markedly from
2014 to 2018, while claims-based performance measures were largely unchanged ....................
8-13 Number of FFS IRF cases increased in 2018 ............................................................................
8-14 Most common types of FFS inpatient rehabilitation facility cases, 2018 ....................................
8-15 Inpatient rehabilitation facilities’ Medicare margins by type of facility, 2010–2018 ...................
ix
8-16 Low standardized costs led to high margins for both hospital-based and freestanding
IRFs, 2018 .............................................................................................................................
8-17 Risk-adjusted quality indicators for IRFs held steady or improved slightly from 2012 to 2018 ...
8-18 The top 25 MS–LTC–DRGs accounted for almost 70 percent of LTCH discharges in 2018 .......
8-19 The number of Medicare LTCH cases and users decreased by over 11 percent between
2017 and 2018 ........................................................................................................................
8-20 The aggregate LTCH Medicare margin increased in 2018 ........................................................
8-21 The share of LTCH cases meeting the criteria for the standard LTCH PPS rate increased
from 2017 to 2018 ..................................................................................................................
9-1 Enrollment in MA plans, 2003–2020 .......................................................................................
9-2 MA plans available to almost all Medicare beneficiaries, 2012–2020 ........................................
9-3 Average monthly rebate dollars, by plan type, 2015–2020 ........................................................
9-4 Changes in enrollment vary among major plan types ................................................................
9-5 MA and cost plan enrollment by state and type of plan, 2020 ...................................................
9-6 MA plan benchmarks, bids, and Medicare program payments relative to FFS spending, 2020 ........
9-7 Enrollment in employer group MA plans, 2007–2020 ..............................................................
9-8 Number of special needs plan enrollees, 2011–2020.................................................................
9-9 Number of SNPs and SNP enrollment rose from 2019 to 2020 .................................................
9-10 The share of Medicare beneficiaries in private plans does not differ substantially in medically
underserved areas compared with other areas, but is somewhat lower in rural areas, 2020 .........
9-11 MA and other private plan enrollment patterns do not differ by medically underserved area
designation but do vary based on urban influence designation, 2020 .........................................
9-12 MA plans are available to nearly all beneficiaries in medically underserved and rural
areas, 2020 .............................................................................................................................
9-13 Most Medicare beneficiaries have access to a considerable number of MA plans, but rural
beneficiaries and beneficiaries in counties composed entirely of MUAs typically have
fewer plans from which to choose, 2020 ..................................................................................
9-14 Twenty most common condition categories among MA beneficiaries, as defined in the
CMS–HCC model, 2018 .........................................................................................................
9-15 Medicare private plan enrollment patterns, by age and Medicare–Medicaid dual-eligible
status, December 2018 ............................................................................................................
10-1 Medicare spending for Part B drugs furnished by physicians, hospital outpatient
departments, and suppliers, 2005–2018 ...................................................................................
10-2 Change in Medicare payments and utilization for separately payable Part B
drugs, 2009–2018 ................................................................................................................
10-3 Top 10 Part B drugs paid based on ASP, by type of provider, 2017 and 2018 ............................
10-4 Growth in ASP for the 20 highest expenditure Part B drugs, 2005–2020 ...................................
10-5 Trends in Medicare Part B payment rates for originator biologics and their biosimilar products .
x
10-6 Price indexes for Medicare Part B drugs, 2005–2018 ...............................................................
10-7 In 2018, 88 percent of Medicare beneficiaries were enrolled in Part D plans or had other
sources of creditable drug coverage .........................................................................................
10-8 Changes in parameters of the Part D defined standard benefit over time ....................................
10-9 Characteristics of stand-alone Medicare PDPs .........................................................................
10-10 Characteristics of MA–PDs .....................................................................................................
10-11 Change in average Part D premiums, 2016–2020 .....................................................................
10-12 More premium-free PDPs for LIS enrollees in 2020 .................................................................
10-13 In 2020, most Part D enrollees are in plans that use a five-tier formulary structure .....................
10-14 In 2020, the share of listed drugs subject to some utilization management increased slightly ......
10-15 Characteristics of Part D enrollees, 2018..................................................................................
10-16 Part D enrollment trends, 2007–2018.......................................................................................
10-17 Part D enrollment by region, 2018 ...........................................................................................
10-18 Components of Part D spending growth...................................................................................
10-19 The majority of Part D spending was incurred by just over one-fifth of all Part D
enrollees, 2018 .......................................................................................................................
10-20 Characteristics of Part D enrollees, by benefit phase reached, 2018 ...........................................
10-21 Part D spending and use per enrollee, 2018 ..............................................................................
10-22 Trends in Part D spending and use per enrollee per month, 2007–2018 .....................................
10-23 Top 15 therapeutic classes of drugs covered under Part D, by spending and volume, 2018 .........
10-24 Part D patterns of prescribing by provider type, 2017 ...............................................................
10-25 Part D patterns of prescribing for selected specialties, 2017 ......................................................
10-26 Price growth for Part D–covered drugs, 2006–2018 .................................................................
10-27 Comparison of price growth for Part B and Part D biologics, 2006–2018 ..................................
11-1 Number of dialysis facilities is growing, and most facilities are for profit and freestanding .............
11-2 Medicare spending for outpatient dialysis services furnished by freestanding and hospital-based
dialysis facilities, 2017 and 2018 .......................................................................................................
11-3 The ESRD population is growing, and most ESRD patients undergo dialysis ..................................
11-4 Asian Americans and Hispanics are among the fastest growing segments of the
ESRD population ...............................................................................................................................
11-5 Characteristics of Medicare fee-for-service dialysis patients, 2018 ...................................................
11-6 Aggregate margins varied by type of freestanding dialysis facility, 2018 .........................................
11-7 Dialysis quality of care: Some measures show progress, others need improvement, 2012–2017 .....
11-8 Hospice spending and use increased in 2018 .....................................................................................
11-9 Hospice use increased across beneficiary groups from 2000 to 2018 ................................................
11-10 Number of Medicare-participating hospices has increased due to growth in for-profit hospices ......
11-11 Hospice cases and length of stay, by diagnosis, 2018 ........................................................................
11-12 Hospice average length of stay among decedents increased slightly in 2018 ....................................
11-13 Hospice length of stay among decedents, by beneficiary and hospice characteristics, 2018 .............
11-14 More than half of Medicare hospice spending in 2018 was for patients with stays exceeding
180 days .............................................................................................................................................
xi
11-15 Hospice aggregate Medicare margins, 2013–2017 ............................................................................
11-16 Medicare margins were higher among hospices with more long stays, 2017 ....................................
11-17 Hospices that exceeded Medicare’s annual payment cap, selected years ..........................................
11-18 Hospice live-discharge rates, 2016–2018...........................................................................................
11-19 Medicare spending for clinical laboratory services, 2005–2019 ........................................................
1
A Data Book: Health care spending and the Medicare program, July 2020 3
Chart 1-1. Medicare was the largest single purchaser of personal health care, 2018
Note: CHIP (Children’s Health Insurance Program), DoD (Department of Defense), VA (Department of Veterans Affairs).
“Personal health care” is a subset of national health expenditures. It includes spending for all medical goods and services that are provided for the treatment of an individual and excludes other spending, such as government administration, the net cost of health insurance, public health, and investment. “Out-of-pocket” spending includes cost sharing for both
privately and publicly insured individuals. Premiums are included in the shares of each program (e.g., Medicare, private health insurance) rather than in the share of the out-of-pocket category. “Other third-party payers” includes worksite health care, other private revenues, Indian Health Service, workers’ compensation, general assistance, maternal and child health, vocational rehabilitation, other federal programs such as the Substance Abuse and Mental Health Services
Administration, other state and local programs, and school health. Slices do not total 100 percent because of rounding. Source: CMS Office of the Actuary, National Health Expenditure Accounts, “Table 6: Personal Health Care Expenditures; Levels,
Percent Change, and Percent Distribution, by Source of Funds: Selected Calendar Years 1970–2018,” released December 2019.
• Medicare is the largest single purchaser of health care in the United States. (Though the share of spending accounted for by private health insurance is greater than Medicare’s share, private health insurance is not a single purchaser of health care; rather, it includes many private plans, including managed care, self-insured health plans, and indemnity plans.) Of the $3.1 trillion spent on personal health care in 2018, Medicare accounted for 23 percent, or $697 billion. This amount includes spending on direct patient care and excludes certain administrative and business costs.
• Thirty-five percent of personal health care spending was financed through private health insurance, and 12 percent was consumer out-of-pocket spending.
• In this chart, Medicare and private health insurance spending include premium contributions from enrollees.
Medicare23%
Medicaid17%Private health
insurance35%
Out of pocket12%
CHIP, DoD, and VA4%
Total = $3.1 trillion
Other third-party payers
8%
4 National health care and Medicare spending
Chart 1-2. Medicare spending is concentrated in certain services and has shifted over time
Total spending 2010 = $517 billion Total spending 2019 = $787 billion
Note: DME (durable medical equipment), SNF (skilled nursing facility). All data are by calendar year. Dollar amounts are Medicare spending only and do not include beneficiary cost sharing. “Other” includes items such as laboratory services, physician-administered drugs, renal dialysis performed in freestanding dialysis facilities, services provided in freestanding
ambulatory surgical center facilities, and ambulance. Components may not total 100 percent because of rounding.
Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020.
• In 2019, Medicare spent $787 billion on benefits. Managed care (Medicare Advantage) was the largest spending category (34 percent), followed by inpatient hospital services (19 percent), prescription drugs provided under Part D (13 percent), and services reimbursed under the physician fee schedule (9 percent).
• The distribution of Medicare spending among services has changed over time. Spending on Medicare Advantage plans has grown from 22 percent of Medicare spending in 2010 to 34 percent in 2019. This growth is largely due to a 96 percent increase in the number of beneficiaries enrolled in Medicare Advantage over this period (data not shown). Meanwhile, the number of beneficiaries in fee-for-service Medicare has stayed relatively flat (data not shown).
• Spending on fee-for-service (FFS) inpatient hospital services has declined as a share of total Medicare spending, falling from 26 percent in 2010 to 19 percent in 2019. Spending on physician fee schedule services has also declined as a share of Medicare spending, falling from 13 percent to 9 percent over this period. At the same time, spending on FFS outpatient services has grown (from 5 percent to 7 percent of Medicare spending), partly due to physician practices being acquired by hospitals and beginning to bill under the outpatient payment system.
Inpatient
hospital19%
Outpatient
hospital7%
Physician fee
schedule9%
Prescription
drugs provided
under Part D
13%SNF
4%Home health
2%
Hospice
3%
DME
1%
Other
8%
Managed
care34%
Inpatient
hospital26%
Outpatient
hospital5%
Physician
fee schedule
13%Prescription
drugs provided under Part D
12%
SNF
5%
Home
health4%
Hospice
3%
DME
2%
Other
8%
Managed
care22%
A Data Book: Health care spending and the Medicare program, July 2020 5
Chart 1-3. Aggregate Medicare spending for FFS beneficiaries, by sector, 2010–2019
Note: FFS (fee-for-service). “Physician fee schedule” includes spending on services provided by physicians and other health professionals such as nurse practitioners, physician assistants, and physical therapists. Dollar amounts are Medicare spending for FFS beneficiaries only and do not include beneficiary cost sharing or spending for Medicare Advantage
enrollees. Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020.
• Medicare fee-for-service spending on inpatient hospital services and physician fee schedule services increased modestly from 2010 to 2019, averaging 1.1 percent and 1.3 percent growth per year, respectively. Spending on skilled nursing facilities and home health services decreased over this period, contracting by –0.2 percent per year on average.
• In contrast, spending on outpatient hospital services doubled during this period (averaging growth of 8.3 percent per year from 2010 to 2019) as more physician practices were acquired by hospitals and began billing Medicare’s outpatient payment system.
133 135 138 140 140 139145 146 147 147
6569 70 70 71 70 70 70 71 73
28 31 34 3641
45 48
5155
58
47 50 46 47 47 48 47
47 46 46
0
20
40
60
80
100
120
140
160
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Med
icare
sp
en
din
g (d
oll
ars
in
bil
lio
ns)
Calendar year
Inpatient hospital
Physician fee schedule
Outpatient hospital
Skilled nursing facilities and home health agencies
6 National health care and Medicare spending
Chart 1-4. Per capita Medicare spending for FFS beneficiaries, by sector, 2010–2019
Note: FFS (fee-for-service), SNF (skilled nursing facility), HH (home health). “Physician fee schedule” includes spending on
services provided by physicians and other health professionals such as nurse practitioners, physician assistants, and physical
therapists. Dollar amounts are Medicare spending for FFS beneficiaries only and do not include beneficiary cost sharing or spending for Medicare Advantage enrollees. Spending per beneficiary for inpatient hospital services equals spending for the sector (see Chart 1-3) divided by FFS enrollment in Medicare Part A. Spending per beneficiary for physician fee schedule
services and outpatient hospital services equals spending for the sector (see Chart 1-3) divided by FFS enrollment in Medicare Part B. Spending per beneficiary for skilled nursing facility services and home health services equals spending for those sectors (see Chart 1-3) divided by total FFS enrollment.
Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020.
• Consistent with the trends shown in Chart 1-3, Medicare FFS spending per capita on inpatient hospital services and physician fee schedule services increased modestly from 2010 to 2019 (averaging 0.4 percent and 1.0 percent per year, respectively). Per capita spending on skilled nursing facilities and home health services decreased over this period (averaging –0.9 percent per year).
• Also consistent with trends in Chart 1-3, per capita spending on outpatient hospital services almost doubled during this period (averaging growth of 7.9 percent per year from 2010 to 2019).
3,737 3,740 3,722 3,745 3,723 3,6923,784 3,795 3,828 3,888
2,032 2,107 2,127 2,100 2,130 2,109 2,078 2,084 2,1152,215
880 948 1,023 1,0861,239
1,344 1,414 1,5211,641
1,743
1,317 1,362 1,244 1,241 1,241
1,259 1,222 1,199 1,192 1,214
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Med
icare
sp
en
din
g (d
oll
ars
per
cap
ita)
Calendar year
Inpatient hospital Physician fee schedule
Outpatient hospital SNF & HH
A Data Book: Health care spending and the Medicare program, July 2020 7
Chart 1-5. Medicare’s share of spending on personal health care varied by type of service, 2018
Note: CCR (continuing care retirement), CHIP (Children’s Health Insurance Program). “Personal health care” is a subset of
national health expenditures. It includes spending for all medical goods and services that are provided for the treatment of an individual and excludes other spending such as government administration, the net cost of health insurance, public health, and investment. “Other” includes private health insurance, out-of-pocket spending, and other private and public
spending. Other service categories included in personal health care that are not shown here are other professional services; dental services; other health, residential, and personal care; and other nondurable medical equipment. Bars may not total 100 percent because of rounding.
Source: CMS Office of the Actuary, National Health Expenditure Accounts, historical data released December 2019.
• While Medicare’s share of total personal health care spending was 23 percent in 2018 (see Chart 1-1), its share of spending by type of service varied, from 16 percent of spending on durable medical equipment to 39 percent of spending on home health and hospice services.
• Medicare’s share of spending on nursing homes and continuing care retirement facilities was smaller than Medicaid’s share. Medicare pays for nursing home services only for Medicare beneficiaries who require skilled nursing or rehabilitation services, whereas Medicaid pays for custodial care (assistance with activities of daily living) provided in nursing homes for people with limited income and assets.
25 23
39
2316
32
1711
35
30
15
11
5865
25
48
69
57
0
10
20
30
40
50
60
70
80
90
100
Hospital Clinician Home healthand hospice
Nursing homeand CCRfacilities
Durablemedical
equipment
Retailprescription
drugs
Sh
are
of
sp
en
din
g (in
perc
en
t)
Medicare Medicaid and CHIP Other
8 National health care and Medicare spending
Chart 1-6. Health care spending has consumed an increasing share of the country’s GDP
Note: GDP (gross domestic product). The potential effects of the COVID-19 pandemic are not reflected in these projections. Source: CMS Office of the Actuary, National Health Expenditure Accounts, historical data released December 2019 and
projections released March 2020.
• In 2018, total health care spending made up 17.7 percent of the country’s GDP. Private health insurance spending constituted 6.0 percent of GDP spending, Medicare constituted 3.6 percent, and Medicaid constituted 2.9 percent.
• Health care spending as a share of GDP more than doubled from 1975 to 2015, increasing from 7.9 percent to 17.6 percent. Private health insurance spending, Medicare spending, and Medicaid all more than tripled over that same time period, increasing from 1.8 percent to 5.8 percent, from 1.0 percent to 3.6 percent, and from 0.8 percent to 3.0 percent, respectively, as a share of GDP.
0
5
10
15
20
25
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Sh
are
of
GD
P (
in p
erc
en
t)
Calendar year
Total health care spending
Private health insurance spending
Medicare spending
Medicaid spending
Historical
Projected
A Data Book: Health care spending and the Medicare program, July 2020 9
Chart 1-7. Trustees project Medicare spending to continue to increase as a share of GDP
Note: GDP (gross domestic product). The Part D benefit began in 2006. Shares for 2025 and later are projections based on the
Trustees’ intermediate set of assumptions. The potential effects of the COVID-19 pandemic are not reflected in these
projections. Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020.
• Over time, Medicare spending has accounted for an increasing share of GDP. From 1 percent in 1975, it is projected to reach 6 percent of GDP in 2045.
• The Medicare Trustees project that spending will rise from 3.6 percent of GDP in 2015 to 5.7 percent of GDP by 2035, largely because of rapid growth in the number of beneficiaries, and then to 6.5 percent of GDP by 2075, with growth in spending per beneficiary becoming the greater factor in the later years of the forecast. The rapid growth in the number of beneficiaries began in 2011 and will continue through 2030 as members of the baby-boom generation reach age 65 and become eligible to enroll in Medicare.
• In the later decades of the Trustees’ forecast, Medicare spending is projected to continue rising as a share of GDP, but at a slower pace than in the past.
• Drug costs are projected to grow faster than Part A and Part B expenditures, and to account for 14 percent of Medicare expenditures by 2085.
1.0
1.7
2.42.6
3.6
4.5
5.7
6.0 6.16.3
6.5 6.5
0
1
2
3
4
5
6
7
1975 1985 1995 2005 2015 2025 2035 2045 2055 2065 2075 2085
Sh
are
of
GD
P (
in p
erc
en
t)
Calendar year
Part DPart BPart A
10 National health care and Medicare spending
Chart 1-8. Per beneficiary spending growth slowed in recent years but is projected to accelerate
Note: CBO (Congressional Budget Office). The potential effects of the COVID-19 pandemic are not reflected in these
projections. Bar totals reflect average annual increase in total Medicare spending (including both fee-for-service and
Medicare Advantage enrollees) and may, because of rounding, differ from the sum of the average annual increase in spending per beneficiary and the average annual increase in Medicare enrollment. Trustees data are presented for calendar years. CBO data are presented for fiscal years.
Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020 and the Congressional Budget Office’s
March 2020 Medicare Baseline.
• The growth in Medicare’s per beneficiary spending slowed in the past decade. The average annual increase equaled or exceeded 6 percent in the 1980s, 1990s, and 2000s, but fell to 2 percent between 2010 and 2019. For 2020 to 2029, the Trustees and CBO project that growth in per beneficiary spending will accelerate but remain lower than historical highs, with the Trustees expecting average annual growth in spending per beneficiary of 5 percent, and the CBO expecting average annual growth of 4 percent.
• The aging of the baby-boom generation accelerated Medicare enrollment growth over the last decade. The average annual growth rate rose to 3 percent between 2010 and 2019. Medicare enrollment is expected to increase an average of 2 percent per year in the next decade.
• Total Medicare spending over the next decade is projected by the Trustees and CBO to increase by an average of 7 percent annually, which would outpace the projected average annual GDP growth of about 4 percent (data not shown).
2 1 23 2 2
10
6
7
2
5 4
0
2
4
6
8
10
12
14
1980s 1990s 2000s 2010–2019 Trustees CBO
Avera
ge a
nn
ual in
cre
ase (
in p
erc
en
t)
Spending per beneficiary
Enrollment
12
9
5
77
7
Historical Projected, 2020−2029
A Data Book: Health care spending and the Medicare program, July 2020 11
Chart 1-9. Employer-sponsored insurance premiums have risen more than twice as fast as Medicare FFS costs
Note: FFS (fee-for-service), HMO (health maintenance organization), PPO (preferred provider organization). The chart shows cumulative growth since 2010. Medicare spending includes Part A and Part B benefits and Part D coverage. Part D spending does not include the portion of premiums paid by enrollees, but does include: (1) Part D spending on non-FFS
beneficiaries enrolled in Medicare Advantage plans with prescription drug coverage, (2) premiums and cost sharing that Medicare pays on behalf of some low-income beneficiaries (who can enroll in either FFS or Medicare Advantage plans), and (3) subsidies Medicare pays employers and unions for drug coverage offered to retirees. Medicare spending includes
the effects of the sequester that began in March 2013, which reduced program spending by 2 percent. Source: Employer-sponsored premium data are from Kaiser Family Foundation surveys, 2010–2019. Medicare spending figures
are from MedPAC analysis of data from the 2020 annual report of the Boards of Trustees of the Medicare trust funds.
• Employer-sponsored insurance premiums have risen faster than the cost of Medicare Part A, Part B, and Part D benefits, despite the richness of employer plans decreasing (due to higher deductibles over time) and the richness of the Medicare benefit increasing (due to changes to Part D). Changes in law have resulted in the phaseout of Part D’s coverage gap—the phase of drug benefit spending in which beneficiaries previously paid much higher cost sharing. Much of the increased generosity was financed by requiring manufacturers of brand-name drugs to discount their prices in the coverage gap.
• Increased prices were largely responsible for spending growth in the private sector. One key driver of the private sector’s higher prices has been provider market power. Hospitals and physician groups have increasingly consolidated, in part to gain leverage over insurers in negotiating higher payment rates. By 2017, 57 percent of hospital markets were so concentrated that one health system produced a majority of hospital discharges (data not shown). Studies have found that prices tend to increase as consolidation increases.
41%
50%
15%
0%
10%
20%
30%
40%
50%
60%
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Perc
en
t
Employer-sponsored HMO premiumgrowth for a single person
Employer-sponsored PPO premiumgrowth for a single person
Medicare per capita FFS spendinggrowth (Part A, Part B, Part D),including the sequester
12 National health care and Medicare spending
Chart 1-10. Trustees and CBO project Medicare spending to exceed $1 trillion by 2022
Note: CBO (Congressional Budget Office). The potential effects of the COVID-19 pandemic are not reflected in these
projections. All data are nominal, mandatory outlays (benefit payments plus mandatory administrative expenses) by fiscal year.
Source: Congressional Budget Office's March 2020 Medicare Baseline; the annual report of the Boards of Trustees of the Medicare trust funds 2020.
• Medicare spending has more than doubled since 2005, increasing from $337 billion to $782 billion by 2019 (these data are by fiscal year and include benefit payments and mandatory administrative expenses).
• The Medicare Trustees and CBO both project that spending for Medicare between 2019 and 2029 will grow at an average annual rate of 6.8 percent. Medicare spending will reach $1 trillion in 2022 under both sets of projections.
• Forecasts of future Medicare spending are inherently uncertain, and differences can stem from different assumptions about the economy that in turn affect annual updates to provider payments and the number of workers paying Medicare payroll taxes. In addition, forecasts can assume different amounts of growth in the volume and intensity of services delivered to Medicare beneficiaries, among other factors.
0
200
400
600
800
1,000
1,200
1,400
1,600
1,800
2005 2010 2015 2020 2025
Do
llars
(in
bil
lio
ns)
Fiscal year
Trustees
CBO
Actual Projected
A Data Book: Health care spending and the Medicare program, July 2020 13
Chart 1-11. FFS program spending was highly concentrated in a small group of beneficiaries, 2017
Note: FFS (fee-for-service). Analysis excludes beneficiaries with any enrollment in a Medicare Advantage plan or other health plan that covers Part A and Part B services (e.g., Medicare cost plans, Medicare–Medicaid Plans, and Medicare and Medicaid’s Program of All-Inclusive Care for the Elderly [PACE]). Components do not sum to totals due to rounding.
Source: Medicare Current Beneficiary Survey, 2017.
• Medicare FFS spending is concentrated among a small number of beneficiaries. In 2017, the costliest 5 percent of beneficiaries (i.e., adding the costliest 1 percent and the next-costliest 4 percent at the top of the bar at left) accounted for 41 percent of annual Medicare FFS spending. The costliest 25 percent of beneficiaries accounted for 83 percent of Medicare spending (calculated on unrounded numbers). The least costly 50 percent of beneficiaries accounted for only 4 percent of FFS spending.
• Costly beneficiaries tend to be those who have multiple chronic conditions, are using inpatient hospital services, are dually eligible for Medicare and Medicaid, and are in the last year of life.
Least costly 50%
4%
Next 25%
13%
Next 15%
25%
Next 5%
18%
Next 4%
25%
Costliest 1%
16%
0
10
20
30
40
50
60
70
80
90
100
Percent of beneficiaries Percent of program spending
Perc
en
t 83%
14 National health care and Medicare spending
Chart 1-12. Medicare HI Trust Fund is projected to be depleted in 2026 under Trustees’ intermediate assumptions
Note: HI (Hospital Insurance). The potential effects of the COVID-19 pandemic are not reflected in these projections. The primary source of income for HI is the payroll tax on covered earnings. Other HI income sources include (1) a portion of the federal income taxes that Social Security recipients with incomes above certain thresholds pay on their benefits and
(2) interest paid on the U.S. Treasury securities held in the HI Trust Fund. *Costs and income for 2019 represent actual (not projected) experience. **Under the low-cost assumption, HI Trust Fund costs would be below income through the 75-year projection period
ending in 2094. Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020.
• The HI Trust Fund, which helps pay for Part A services such as inpatient hospital stays and post-acute care provided by skilled nursing facilities and hospice, is mainly financed through a dedicated payroll tax (i.e., a tax on wage earnings).
• From 2008 to 2015, the HI Trust Fund ran an annual deficit (i.e., paid more in benefits than it collected in payroll taxes) (data not shown). In 2016 and 2017, the HI Trust Fund ran a surplus (data not shown). However, deficits returned in 2018 and 2019 and are projected to continue until trust fund assets are depleted in 2026 (under the Trustees’ intermediate assumptions). Under high-cost assumptions, the HI Trust Fund could be depleted as early as 2023. Under low-cost assumptions, it would remain able to pay full benefits indefinitely.
• The Trustees estimate that the payroll tax would need to be immediately increased from its current rate of 2.90 percent to 3.66 percent to balance the HI Trust Fund over the next 75 years. Alternatively, Part A spending would need to be immediately reduced by 16 percent (data not shown).
Costs below income Costs exceed income HI Trust Fund assets exhausted
High cost
Intermediate
Low cost
2019* 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 …………….…………….…………….…………….…………….2094**
Calendar year
Esti
mate
A Data Book: Health care spending and the Medicare program, July 2020 15
0%
1%
2%
3%
4%
5%
6%
7%
1966 1976 1986 1996 2006 2016 2026 2036 2046 2056 2066 2076
Sh
are
of
GD
P
Calendar year
Historical ProjectedTotal Medicare spending
General revenue transfers
State transfers and drug fees
Premiums
Payroll taxes
Tax on benefits
Part A deficit
Chart 1-13. General revenues have overtaken Medicare payroll taxes as the largest source of Medicare funding
Note: GDP (gross domestic product). These projections are based on the Trustees’ intermediate set of assumptions and do not reflect the potential effects of the COVID-19 pandemic. “Tax on benefits” refers to the portion of income taxes that higher income individuals pay on Social Security benefits, which is designated for Medicare. “State transfers” (often called the
Part D “clawback”) refers to payments from the states to Medicare for assuming primary responsibility for prescription drug spending that were mandated by the Medicare Prescription Drug, Improvement, and Modernization Act of 2003. “Drug fees” refers to the fee imposed by the Patient Protection and Affordable Care Act of 2010 on manufacturers and
importers of brand-name prescription drugs. These fees are deposited in the Part B account of the Supplementary Medical Insurance Trust Fund.
Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020.
• Medicare spending accounted for 3.7 percent of GDP in 2019. The Medicare Trustees project that Medicare’s share of GDP will rise to 5.5 percent by 2033 and to 5.9 percent by 2038.
• In the early years of the Medicare program, payroll taxes deposited into Medicare’s Hospital Insurance Trust Fund (which finances Part A) were the main source of funding for the program, but beginning in 2009, general revenue transfers (which help finance Part B and Part D) became the largest single source of Medicare income. General revenue transfers are expected to continue to be a substantial share of Medicare financing, growing to about 49 percent by 2034, and then remaining stable through the rest of the century.
• As more general revenues are devoted to Medicare, fewer resources will be available to invest in growing the economic output of the future or in supporting other national priorities.
Dedicated
revenues
2086
16 National health care and Medicare spending
Chart 1-14. Medicare enrollment is rising while the number of workers per HI beneficiary is declining
Medicare HI enrollment Workers per HI beneficiary
Note: HI (Hospital Insurance). Hospital Insurance is also known as Medicare Part A. The potential effects of the COVID-19
pandemic are not included in these projections. Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020.
• As the baby-boom generation ages, enrollment in the Medicare program is surging. By 2033, Medicare is projected to have over 80 million beneficiaries—up from 62 million beneficiaries in 2020.
• While Medicare enrollment is rising, the number of workers per beneficiary is rapidly declining. Workers are the primary funder of Medicare’s HI Trust Fund, which they fund through payroll taxes. However, the number of workers per Medicare beneficiary has declined from 4.6 during the early years of the program to 2.9 in 2020 and is projected by the Medicare Trustees to fall to 2.5 by 2028.
• These demographics threaten the financial stability of the Medicare program.
0
20
40
60
80
100
1970 1990 2010 2030
En
roll
men
t (i
n m
illi
on
s)
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1970 1990 2010 2030
Wo
rkers
per
HI b
en
efi
cia
ry
A Data Book: Health care spending and the Medicare program, July 2020 17
Chart 1-15. Medicare HI and SMI benefits and cost sharing per FFS beneficiary, 2018
Average benefit in 2018
(in dollars) Average cost sharing in 2018
(in dollars)
HI (Part A) $4,972 $415
SMI (Part B, excludes Part D) 5,959 1,513
Note: HI (Hospital Insurance), SMI (Supplementary Medical Insurance), FFS (fee-for-service). Dollar amounts are nominal for
FFS Medicare only and do not include Part D. “Average benefit” represents amounts paid for covered services per FFS
beneficiary and excludes administrative expenses. “Average cost sharing” represents the sum of deductibles, coinsurance, and balance billing paid for covered services per FFS beneficiary and excludes all monthly premiums.
Source: CMS Program Statistics, CMS Office of Enterprise Data and Analytics, CMS Chronic Conditions Data Warehouse.
• In calendar year 2018, the Medicare program made $4,972 in HI (Part A) benefit payments and $5,959 in SMI (Part B) benefit payments on average per FFS beneficiary.
• Beneficiaries owed an average of $415 in cost sharing for HI and $1,513 in cost sharing for SMI in calendar year 2018. (Cost sharing excludes all monthly premiums.)
• To cover some of those cost-sharing requirements, 89 percent of beneficiaries have coverage that supplements or replaces the Medicare benefit package, such as Medicare Advantage, Medicaid, supplemental coverage through former employers, and Medigap coverage.
2
A Data Book: Health care spending and the Medicare program, July 2020 21
Chart 2-1. Aged beneficiaries accounted for the greatest share of the Medicare population and program spending, 2017
Note: ESRD (end-stage renal disease). The “aged” category includes beneficiaries ages 65 and older without ESRD. The “disabled" category includes beneficiaries under age 65 without ESRD. The “ESRD” category includes beneficiaries with ESRD, regardless of age. Results include fee-for-service, Medicare Advantage, community-dwelling, and institutionalized
beneficiaries. The Medicare Current Beneficiary Survey is a point-in-time survey from a sample of Medicare beneficiaries. Year-to-year variation in reported data is expected.
Source: MedPAC analysis of the Medicare Current Beneficiary Survey, Cost Supplement file 2017.
• In 2017, beneficiaries ages 65 and older without ESRD composed 84.6 percent of the beneficiary population and accounted for 77.3 percent of Medicare spending. Beneficiaries under 65 with a disability and beneficiaries with ESRD accounted for the remaining population and spending.
• A disproportionate share of Medicare expenditures is devoted to Medicare beneficiaries with ESRD. On average, these beneficiaries incur spending that is more than five times greater than spending for aged beneficiaries (ages 65 years and older without ESRD) and more than three times greater than spending for beneficiaries under age 65 with a disability (non-ESRD). In 2017, $54,905 was spent per ESRD beneficiary versus $10,978 per aged beneficiary and $15,529 per beneficiary under age 65 enrolled because of disability (data not shown).
77.3%
84.6%
18.8%
14.5%
3.9%
0.9%
Share ofspending
Share ofbeneficiaries
Aged Disabled ESRD
Total beneficiaries: 59.3 million
Total spending: $713 billion
22 Medicare beneficiary demographics
Chart 2-2. Beneficiaries younger than 65 accounted for a disproportionate share of Medicare spending, 2017
Note: Results include fee-for-service, Medicare Advantage, community-dwelling, and institutionalized beneficiaries. The
Medicare Current Beneficiary Survey is a point-in-time survey from a sample of Medicare beneficiaries. Year-to-year variation in reported data is expected.
Source: MedPAC analysis of the Medicare Current Beneficiary Survey, Cost Supplement file 2017.
• Beneficiaries younger than 65 made up 14.8 percent of the beneficiary population in 2017 but accounted for 19.5 percent of Medicare spending.
• In 2017, average Medicare spending per beneficiary was $12,018.
• For the aged population (65 and older), per capita expenditures increase with age. In 2017, per capita expenditures were $9,314 for beneficiaries 65 to 74 years old, $13,194 for those 75 to 84 years old, and $15,959 for those 85 or older (data not shown).
• In 2017, per capita expenditures for Medicare beneficiaries under age 65 who were enrolled because of end-stage renal disease or disability were $15,879 (data not shown).
19.5%
14.8%
37.7%
48.7%
27.6%
25.1%
15.2%
11.4%
Share ofspending
Share ofbeneficiaries
Under 65 65-74 75-84 85+
Average per capita spending = $12,018
A Data Book: Health care spending and the Medicare program, July 2020 23
Chart 2-3. Beneficiaries who reported being in poor health accounted for a disproportionate share of
Medicare spending, 2017
Note: Results include fee-for-service, Medicare Advantage, community-dwelling, and institutionalized beneficiaries. Totals may
not sum to 100 percent due to rounding. “Other” category excluded. The Medicare Current Beneficiary Survey is a point-
in-time survey from a sample of Medicare beneficiaries. Year-to-year variation in reported data is expected. Source: MedPAC analysis of the Medicare Current Beneficiary Survey, Cost Supplement file 2017.
• In 2017, most beneficiaries reported fair to excellent health. Only about 6 percent reported poor health.
• Medicare spending is strongly associated with self-reported health status. In 2016, per capita expenditures were $6,873 for those who reported excellent or very good health, $13,649 for those who reported good or fair health, and $29,109 for those who reported poor health (data not shown).
28.0%
46.5%
57.2%
47.8%
14.7%
5.8%
Share ofspending
Share ofbeneficiaries
Excellent or very good health Good or fair health Poor health
Average per capita spending = $12,018
24 Medicare beneficiary demographics
Chart 2-4. Enrollment in the Medicare program is projected to grow rapidly through 2030
Note: Enrollment numbers are based on Part A enrollment only. Beneficiaries enrolled only in Part B are not included. The
potential effects of the COVID-19 pandemic are not reflected in these projections. Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020.
• The total number of people enrolled in the Medicare program is projected to increase from about 62 million in 2020 to about 78 million in 2030.
• The rate of increase in Medicare enrollment will accelerate until 2030 as more members of the baby-boom generation become eligible, at which point it will continue to increase, but more slowly, after the entire baby-boom generation has become eligible.
20.1
28.0
33.7
39.3
47.4
62.3
77.5
84.087.2
93.2
100.0
106.1109.3
0
20
40
60
80
100
120
1970 1980 1990 2000 2010 2020 2030 2040 2050 2060 2070 2080 2090
Ben
efi
cia
ries (
in m
illi
on
s)
Historic Projected
A Data Book: Health care spending and the Medicare program, July 2020 25
Chart 2-5. Characteristics of the Medicare population, 2017 Share of the Share of the Medicare Medicare Characteristic population Characteristic population Total (54.5 million) 100% Living arrangement Institution 3% Sex Alone 27 Male 46 With spouse 48 Female 54 Other 22 Race/ethnicity Education White, non-Hispanic 75 No high school diploma 15 African American, High school diploma only 26 non-Hispanic 10 Some college or more 57 Hispanic 8 Other 7 Income status Below poverty 15 Age 100–125% of poverty 7 <65 15 125–200% of poverty 17 65–74 47 200–400% of poverty 27 75–84 26 Over 400% of poverty 35 85+ 11 Supplemental insurance status Health status Medicare only 16 Excellent or very good 46 Managed care 33 Good or fair 47 Employer-sponsored insurance 19 Poor 7 Medigap 19 Medigap with employer- Residence sponsored insurance 1 Urban 80 Medicaid 11 Rural 20 Other 1 Note: Total number of beneficiaries, age, and health status values may slightly differ from previous figures because only beneficiaries
with complete characteristic data were included in this analysis. Totals may not sum to 100 percent due to rounding and exclusion of an “other” category. “Urban” indicates beneficiaries living in metropolitan statistical areas (MSAs). “Rural” indicates beneficiaries living outside MSAs. In 2017, “poverty” was defined as income of $11,756 for single individuals ages 65 and older
and $14,828 for married couples ages 65 and older. Poverty thresholds are calculated by the U.S. Census Bureau (https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-poverty-thresholds.html). Some beneficiaries may have more than one type of supplemental insurance. The Medicare Current Beneficiary Survey is a point-in-time survey
from a sample of Medicare beneficiaries. Year-to-year variation in reported data is expected. Source: MedPAC analysis of the Medicare Current Beneficiary Survey Cost Supplement file 2017.
• The majority of Medicare beneficiaries are female (rather than male) and White (rather than other races/ethnicities).
• About one-fifth of beneficiaries live in rural areas.
• Twenty-seven percent of the Medicare population lives alone.
• Most Medicare beneficiaries have some source of supplemental insurance. Managed care plans are the most common source of supplemental coverage.
3
A Data Book: Health care spending and the Medicare program, July 2020 29
Chart 3-1. Sources of supplemental coverage among noninstitutionalized Medicare beneficiaries, 2017
Note: We assigned beneficiaries to the supplemental coverage category they were in for the most time in 2017. They could have had coverage in other categories during 2017. “Other public sector” includes federal and state programs not included in other categories. Analysis includes only beneficiaries not living in institutions such as nursing homes. It
excludes beneficiaries who were not in both Part A and Part B throughout their enrollment in 2017 or who had Medicare as a secondary payer. Numbers do not total 100 because of rounding. CMS adjusted the beneficiary weights used in the Medicare Current Beneficiary Survey for 2017 so that the estimated number of beneficiaries in the Medicare Advantage
program matched a control total. Differences between this chart and those published in previous Data Books may not reflect real change but rather may be due to differences in method.
Source: MedPAC analysis of Medicare Current Beneficiary Survey, Survey file 2017.
• Most beneficiaries living in the community (noninstitutionalized) have coverage that supplements or replaces the Medicare benefit package. In 2017, 89 percent of beneficiaries had supplemental coverage or participated in Medicare managed care.
• About 41 percent of beneficiaries had private sector supplemental coverage such as Medigap (about 22 percent) or employer-sponsored retiree coverage (about 19 percent).
• About 11 percent of beneficiaries had public sector supplemental coverage, primarily Medicaid.
• About 37 percent of beneficiaries participated in Medicare managed care. This care includes Medicare Advantage, health care prepayment, and cost plans. These types of arrangements generally replace Medicare’s fee-for-service coverage and often add more coverage.
• The numbers in this chart differ from those in Chart 2-5, Chart 4-1, and Chart 4-4 because of differences in the populations represented in the charts. This chart excludes beneficiaries in long-term care institutions, while Chart 2-5 and Chart 4-4 include all Medicare beneficiaries, and Chart 4-1 excludes beneficiaries in Medicare Advantage.
Medigap21.9%
Employer-sponsored insurance
19.1%
Medicaid10.5%
Other public sector0.5%
Medicare managed care
36.7%
No supplemental coverage
11.2%
30 Medicare beneficiary and other payer financial liability
Chart 3-2. Sources of supplemental coverage among noninstitutionalized Medicare beneficiaries, by beneficiaries’ characteristics, 2017
Number of Employer- Medicare Other beneficiaries sponsored Medigap managed public Medicare (thousands) insurance insurance Medicaid care sector only
All beneficiaries 47,364 19% 22% 11% 37% 1% 11%
Age <65 7,048 6 4 38 35 1 16 65–69 10,557 19 22 7 38 0 14 70–74 11,239 22 27 5 36 1 9 75–79 8,018 21 26 5 38 0 9 80–84 5,322 23 25 6 38 0 9 85+ 5,180 24 24 6 36 0 9 Income-to-poverty ratio ≤1.00 7,986 4 7 42 35 1 10 1.00 to 1.20 2,842 7 11 24 43 1 14 1.20 to 1.35 2,207 8 16 13 43 1 19 1.35 to 2.00 7,670 14 22 6 42 1 15 >2.00 26,659 29 28 1 34 0 9 Eligibility status Aged 40,098 21 25 6 37 0 10
Disabled 6,839 6 4 38 35 1 16 ESRD 372 14 23 24 19 1 20 Residence Urban 37,627 19 20 10 40 1 10 Rural 9,736 18 28 14 24 0 16 Sex Male 21,312 20 21 10 35 1 14 Female 26,052 18 23 11 38 0 9 Health status Excellent/very good 21,771 23 26 4 37 0 10 Good/fair 22,340 17 19 14 37 1 12 Poor 3,045 8 14 28 33 1 15
Note: ESRD (end-stage renal disease). We assigned beneficiaries to the supplemental coverage category they were in for the most time in 2017. They could have had coverage in other categories during 2017. “Medicare managed care” includes
Medicare Advantage, cost, and health care prepayment plans. “Other public sector” includes federal and state programs not included in other categories. “Urban” indicates beneficiaries living in metropolitan statistical areas (MSAs) as indicated by core-based statistical areas. “Rural” indicates beneficiaries living outside MSAs, which includes both micropolitan
statistical areas and rural areas as indicated by core-based statistical areas. Analysis excludes beneficiaries living in institutions such as nursing homes. It excludes beneficiaries who were not in both Part A and Part B throughout their enrollment in 2017 or who had Medicare as a secondary payer. The number of beneficiaries differs among boldface
categories because we excluded beneficiaries with missing values. Numbers in some rows do not sum to 100 percent because of rounding.
Source: MedPAC analysis of Medicare Current Beneficiary Survey, Survey file 2017.
• Beneficiaries most likely to have employer-sponsored supplemental coverage are those who are age 65 or older, have income above twice the poverty level, are eligible because of age, and report better than poor health.
• Medigap is most common among those who are age 65 or older, have income higher than 1.35 times the poverty level, are eligible because of age or ESRD, are rural dwelling, and report better than poor health.
• Medicaid coverage is most common among those who are under age 65, have income lower than 1.2 times the poverty level, are eligible because of disability or ESRD, are rural dwelling, and report poor health.
• Lack of supplemental coverage (Medicare coverage only) is most common among beneficiaries who are under age 70, are eligible because of disability or ESRD, are rural dwelling, are male, and report poor health.
A Data Book: Health care spending and the Medicare program, July 2020 31
Chart 3-3. Covered benefits and enrollment in standardized Medigap plans, 2018
Medigap standardized plan type
High
deductible
Benefit A B C* D F* F G K L M N
Part A hospital costs ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Part B cost sharing ✓ ✓ ✓ ✓ ✓ ✓ ✓ 50% 75% ✓ $20/$50
Blood (first 3 pints) ✓ ✓ ✓ ✓ ✓ ✓ ✓ 50% 75% ✓ ✓ Hospice cost sharing ✓ ✓ ✓ ✓ ✓ ✓ ✓ 50% 75% ✓ ✓
SNF coinsurance ✓ ✓ ✓ ✓ ✓ 50% 75% ✓ ✓
Part A deductible ✓ ✓ ✓ ✓ ✓ ✓ 50% 75% 50% ✓
Part B deductible ✓ ✓ ✓ Part B excess charges ✓ ✓ ✓ Foreign travel emergency ✓ ✓ ✓ ✓ ✓ ✓ ✓
Lives covered (in thousands) 125 225 700 150 6,750 275 2,300 75 50 5 1,350
Percent change 2016–2018 –20% –17% –21% –18% 1% 18% 82% 8% 0% –13% 17%
Note: SNF (skilled nursing facility). Three states (Massachusetts, Minnesota, and Wisconsin) have different plan types and are
not included in this chart. The ✓ indicates that the plan covers all cost sharing. Percentages indicate that the plan covers
that share of the total cost sharing. The $20/$50 indicates that the plan covers all but $20 for physician office visits and all but $50 for emergency room visits.
*Beginning in 2020, neither the C plan nor the F plan are allowed to cover the Part B deductible for new policies sold.
However, C plans and F plans sold before 2020 can continue to cover the Part B deductible. Source: MedPAC analysis of National Association of Insurance Commissioners data, 2019.
• Medicare beneficiaries often purchase Medigap plans, also known as Medicare supplementary insurance plans, to cover fee-for-service Medicare cost sharing. Statute specifies 11 standardized plans. States enforce the standards based on model regulations developed by the National Association of Insurance Commissioners. Three states (Massachusetts, Minnesota, and Wisconsin) have waivers from these standards and have different standard plan types not included in this chart.
• Plan F, which covers all Medicare cost sharing, is the most popular plan, with 6.8 million enrollees. However, because the Congress was concerned about the overutilization of Medicare services, legislation prohibits the sale of new Plan F policies as of 2020. As a result, insurers have begun to direct beneficiaries into other plan types, namely plans G, K, and N, which do not cover the Part B deductible.
• During 2018, almost 14 million beneficiaries enrolled in Medigap plans (including those in Massachusetts, Minnesota, and Wisconsin). Of all Medicare beneficiaries, about one-fifth were enrolled in Medigap plans.
32 Medicare beneficiary and other payer financial liability
Chart 3-4. Total spending on health care services for noninstitutionalized FFS Medicare beneficiaries, by source of payment, 2017
Note: FFS (fee-for-service). “Private supplements” includes employer-sponsored plans and individually purchased coverage.
“Public supplements” includes Medicaid, Department of Veterans Affairs, and other public coverage. “Beneficiaries’ direct spending” is on Medicare cost sharing and noncovered services, but not supplemental premiums. Analysis includes only FFS beneficiaries not living in institutions such as nursing homes.
Source: MedPAC analysis of Medicare Current Beneficiary Survey, Cost Supplement file, 2017.
• Among FFS beneficiaries living in the community (noninstitutionalized), the total cost of health care services (beneficiaries’ direct spending as well as expenditures by Medicare, other public sector sources, and all private sector sources on all health care goods and services) averaged about $15,800 in 2017. Medicare was the largest source of payment: It paid about 68 percent of the health care costs for FFS beneficiaries living in the community, an average of $10,755 per beneficiary. The level of Medicare spending in this chart differs from the level in Chart 2-1 because this chart excludes beneficiaries in Medicare Advantage and those living in institutions, while Chart 2-1 represents all Medicare beneficiaries.
• Private sources of supplemental coverage—primarily employer-sponsored retiree coverage and Medigap—paid about 13 percent of beneficiaries’ costs, an average of $2,087 per beneficiary.
• Beneficiaries paid about 14 percent of their health care costs out of pocket, an average of $2,163 per beneficiary.
• Public sources of supplemental coverage—primarily Medicaid—paid about 5 percent of beneficiaries’ health care costs, an average of $846 per beneficiary.
Medicare68%
Beneficiaries' direct spending
14%
Private supplements
13%
Public supplements
5%
Per capita total spending = $15,850
A Data Book: Health care spending and the Medicare program, July 2020 33
Chart 3-5. Per capita total spending on health care services among noninstitutionalized FFS beneficiaries, by source of payment, 2017
Note: FFS (fee-for-service). Analysis excludes those who are not in FFS Medicare and those living in institutions such as
nursing homes. “Out-of-pocket" spending includes Medicare cost sharing and noncovered services, but not supplemental premiums.
Source: MedPAC analysis of the Medicare Current Beneficiary Survey, Cost Supplement file, 2017.
• Total spending on health care services varied dramatically among FFS beneficiaries living in the community in 2017. Per capita spending for the 10 percent of beneficiaries with the highest total spending averaged $78,500. Per capita spending for the 10 percent of beneficiaries with the lowest total spending averaged $323.
• Among FFS beneficiaries living in the community, Medicare paid a larger share as total spending increased, and beneficiaries’ out-of-pocket spending was a smaller share as total spending increased. For example, Medicare paid 68 percent of total spending for all beneficiaries, but paid 77 percent of total spending for the 10 percent of beneficiaries with the highest total spending. Beneficiaries’ out-of-pocket spending covered 14 percent of total spending for all beneficiaries, but only 7 percent of total spending for the 10 percent of beneficiaries with the highest total spending (data not shown).
$323 $1,652$4,435
$10,910
$25,785
$78,500
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
90,000
<10 10–25 25–50 50–75 75–90 >90
Do
llars
Groups of beneficiaries ranked by total spending (percentile ranges)
Medicare
Supplemental payers
Out of pocket
4
A Data Book: Health care spending and the Medicare program, July 2020 37
Chart 4-1. Dual-eligible beneficiaries accounted for a disproportionate share of Medicare spending, 2017
Share of FFS beneficiaries Share of FFS spending
Note: FFS (fee-for-service). “Dual-eligible beneficiaries” are defined as beneficiaries who were eligible for both Medicare and
Medicaid for at least one month during the year.
Source: MedPAC analysis of the Medicare Current Beneficiary Survey, 2017.
• Dual-eligible beneficiaries are those who qualify for both Medicare and Medicaid. Medicaid is a joint federal and state program designed to help people with low incomes obtain needed health care.
• Dual-eligible beneficiaries account for a disproportionate share of Medicare FFS expenditures. Although they were 17 percent of the Medicare FFS population in 2017, they represented 30 percent of aggregate Medicare FFS spending.
• On average, Medicare FFS per capita spending is more than twice as high for dual-eligible beneficiaries compared with non-dual-eligible beneficiaries: In 2017, $19,846 was spent per dual-eligible beneficiary, and $9,415 was spent per non-dual-eligible beneficiary (data not shown).
• In 2017, average total spending⎯which includes Medicare, Medicaid, supplemental
insurance, and out-of-pocket spending across all payers⎯for dual-eligible beneficiaries was $30,510 per beneficiary, about twice the amount for other Medicare beneficiaries (data not shown).
Dual eligible17%
Non-dual eligible83%
Dual eligible30%
Non-dual eligible70%
38 Dual-eligible beneficiaries
Chart 4-2. Dual-eligible beneficiaries were more likely than non-dual-eligible beneficiaries to be under age 65 and disabled, 2017
Dual-eligible beneficiaries Non-dual-eligible beneficiaries
Note: Beneficiaries who are under age 65 generally qualify for Medicare because they are disabled. Once disabled beneficiaries reach age 65, they are counted as aged beneficiaries. “Dual-eligible beneficiaries” are defined as beneficiaries who were eligible for both Medicare and Medicaid for at least one month during the year.
Source: MedPAC analysis of Medicare Current Beneficiary Survey, 2017.
• Disability is a pathway for individuals to become eligible for both Medicare and Medicaid benefits.
• Dual-eligible beneficiaries are more likely than non-dual-eligible beneficiaries to be under age 65 and disabled. In 2017, 42 percent of dual-eligible beneficiaries were under age 65 and disabled compared with 9 percent of the non-dual-eligible population.
Under 65 (disabled)
42%
65-7429%
75-8419%
85+10%
Under 65 (disabled)
9%
65-7452%
75-8428%
85+11%
A Data Book: Health care spending and the Medicare program, July 2020 39
Chart 4-3. Dual-eligible beneficiaries were more likely than non-dual-eligible beneficiaries to report being in poor health, 2017
Dual-eligible beneficiaries Non-dual-eligible beneficiaries
Note: “Dual-eligible beneficiaries” are defined as beneficiaries who were eligible for both Medicare and Medicaid for at least one
month during the year. Percentages in the non-dual-beneficiaries pie chart do not total 100 because of rounding.
Source: MedPAC analysis of the Medicare Current Beneficiary Survey, 2017.
• Dual-eligible beneficiaries are more likely than non-dual-eligible beneficiaries to report being in poor health. In 2017, 16 percent of dual-eligible beneficiaries reported being in poor health compared with 5 percent of non-dual-eligible beneficiaries.
• Just over half of non-dual-eligible beneficiaries (52 percent) reported being in excellent or very good health in 2017. In comparison, about one-fifth (21 percent) of dual-eligible beneficiaries reported being in excellent or very good health.
Poor health16%
Good or fair health
63%
Excellent or very good health21%
Poor health
5%
Good or fair health
44%
Excellent or very good health52%
40 Dual-eligible beneficiaries
Chart 4-4. Demographic differences between dual-eligible beneficiaries and non-dual-eligible beneficiaries, 2017
Percent of dual- Percent of non-dual- Characteristic eligible beneficiaries eligible beneficiaries
Sex Male 39% 47% Female 61 53 Race/ethnicity White, non-Hispanic 52 81 African American, non-Hispanic 20 7 Hispanic 18 6 Other 9 7 Limitations in ADLs No limitations in ADLs 47 74 Limitations in 1–2 ADLs 25 17 Limitations in 3–6 ADLs 29 8 Residence Urban 78 80 Rural 22 20 Living arrangement Institution 10 1 Alone 32 26 With spouse 15 56 With children, nonrelatives, others 43 17 Education No high school diploma 38 11 High school diploma only 31 25 Some college or more 30 64 Income status Below poverty 55 5 100–125% of poverty 20 4 125–200% of poverty 17 17 200–400% of poverty 6 32 Over 400% of poverty 1 42 Supplemental insurance status Medicare or Medicare/Medicaid only 55 19 Medicare managed care 38 33 Employer-sponsored insurance 1 23 Medigap 3 23 Medigap/employer <1 1 Other* 3 1
Note: ADL (activity of daily living). “Dual-eligible beneficiaries” are defined as beneficiaries who were eligible for both Medicare and Medicaid for at least one month during the year. “Urban” indicates beneficiaries living in metropolitan statistical areas (MSAs). “Rural” indicates beneficiaries living outside of MSAs. In 2017, poverty was defined as annual income of $11,756 for people living
alone and $14,828 for married couples. Totals may not sum to 100 percent due to rounding and exclusion of an “other” category. Poverty thresholds are calculated by the U.S. Census Bureau (https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-poverty-thresholds.html).
*Includes public programs such as the Department of Veterans Affairs and state-sponsored drug plans.
Source: MedPAC analysis of Medicare Current Beneficiary Survey, 2017.
• Dual-eligible beneficiaries qualify for Medicaid due in part to low incomes. In 2017, 55 percent of dual-eligible beneficiaries lived below the poverty threshold, and 92 percent lived below 200 percent of the poverty threshold. Compared with non-dual-eligible beneficiaries, dual-eligible beneficiaries are more likely to be female, be African American or Hispanic, lack a high school diploma, have greater limitations in activities of daily living, and live in an institution. They are less likely to have supplemental employer-sponsored or Medigap coverage.
A Data Book: Health care spending and the Medicare program, July 2020 41
Chart 4-5. Differences in Medicare spending and service use between dual-eligible beneficiaries and non-dual-eligible beneficiaries, 2017
Dual-eligible Non-dual-eligible Service beneficiaries beneficiaries
Average FFS Medicare payment for all beneficiaries
Total Medicare FFS payments $19,846 $9,415
Inpatient hospital 4,318 2,746
Physiciana 3,410 2,663
Outpatient hospital 2,416 1,484 Home health 744 354
Skilled nursing facilityb 1,397 462
Hospice 509 255
Prescribed medicationc 7,013 1,449
Share of FFS beneficiaries using service
Share using any type of service 93.8% 87.1%
Inpatient hospital 22.5 14.0
Physiciana 89.8 81.9
Outpatient hospital 76.0 62.5 Home health 13.1 7.5
Skilled nursing facilityb 7.4 3.4
Hospice 3.8 2.1
Prescribed medicationc 75.9 59.9
Note: FFS (fee-for-service). Data in this analysis are restricted to beneficiaries in FFS Medicare. “Dual-eligible beneficiaries” are
defined as beneficiaries who were eligible for both Medicare and Medicaid for at least one month during the year. Spending totals derived from the Medicare Current Beneficiary Survey (MCBS) do not necessarily match official estimates from CMS Office of the Actuary. Total payments may not equal the sum of line items due to omitted “other” category.
a Includes a variety of medical services, equipment, and supplies. b Individual short-term facility (usually skilled nursing facility) stays for the MCBS population. c Data from stand-alone prescription drug plans and Medicare Advantage–Prescription Drug plans.
Source: MedPAC analysis of the Medicare Current Beneficiary Survey, 2017.
• In 2017, average per capita Medicare FFS spending for dual-eligible beneficiaries was more
than twice that for non-dual-eligible beneficiaries⎯$19,846 compared with $9,415.
• For each type of service, average Medicare FFS per capita spending was higher for dual-eligible beneficiaries than for non-dual-eligible beneficiaries.
• Higher average per capita FFS spending for dual-eligible beneficiaries is a function of higher use of these services by dual-eligible beneficiaries compared with their non-dual-eligible counterparts. Dual-eligible beneficiaries are more likely than non-dual-eligible beneficiaries to use each type of Medicare-covered service.
42 Dual-eligible beneficiaries
Chart 4-6. Both Medicare and total spending were concentrated among dual-eligible beneficiaries, 2017
Note: “Total spending” includes Medicare, Medicaid, supplemental insurance, and out-of-pocket spending. Data in this analysis
are restricted to beneficiaries in fee-for-service (FFS) Medicare. “Dual-eligible beneficiaries” are defined as beneficiaries
who were eligible for both Medicare and Medicaid for at least one month during the year. Source: MedPAC analysis of the Medicare Current Beneficiary Survey, 2017.
• Annual Medicare FFS and total spending on dual-eligible beneficiaries are concentrated among a small number. The costliest 5 percent of dual-eligible beneficiaries accounted for 35 percent of Medicare spending and 26 percent of total spending on dual-eligible beneficiaries in 2017. In contrast, the least costly 50 percent of dual-eligible beneficiaries accounted for only 6 percent of Medicare FFS spending and 19 percent of total spending on dual-eligible beneficiaries.
• On average, total spending (including Medicaid, Medigap, etc.) for dual-eligible beneficiaries in 2017 was about twice that for non-dual-eligible beneficiaries—$30,510 compared with $15,630, respectively (data not shown).
6
50
19
23
30
25
36
15
30
35
5
26
0
10
20
30
40
50
60
70
80
90
100
Medicare spending for dual-eligible beneficiaries
Share of dual-eligiblebeneficiaries
Total spending for dual-eligiblebeneficiaries
Perc
en
t
5
A Data Book: Health care spending and the Medicare program, July 2020 45
Chart 5-1. Most beneficiaries are in Medicare managed care plans or are assigned to accountable care organizations, 2020
Note: ACO (accountable care organization), FFS (fee-for-service), MSSP (Medicare Shared Savings Program). This chart includes
only beneficiaries enrolled in both Part A and Part B in January 2020. Both Part A and Part B coverage is necessary for either
Medicare Advantage enrollment or ACO assignment. Percentages in this chart differ from those in Charts 9-5 and 9-10 because the 5.7 million beneficiaries with either only Part A or only Part B coverage are included in those charts. In general, Medicare managed care plans include Medicare Advantage plans as well as cost-reimbursed plans. Other ACOs and ACO-
like models include the Next Generation ACO model, the Maryland Total Cost of Care (TCOC) model, the ESRD Seamless Care Organization (ESCO) model, and the Vermont All-Payer ACO. In the Maryland TCOC model, all FFS beneficiaries are assigned to a hospital, and each hospital is responsible for all Part A and Part B spending for all Medicare beneficiaries in its
market. This system creates ACO-like incentives for the hospital and qualifies physicians affiliated with those hospitals for the Medicare Access and CHIP Reauthorization Act (MACRA) bonus payments for participation in eligible alternative payment models. Percentages do not total 100 because of rounding.
Source: CMS January 2020 enrollment dashboard data, CMS Shared Savings Program January 2020 Fast Facts, CMS ACO Next
Generation 2018 performance data and 2019 participant lists, CMS ESCO 2018 report to the Congress, and State of
Vermont Green Mountain Care Board 2020 report.
• Among the 56.5 million Medicare beneficiaries with both Part A and Part B coverage in 2020, approximately two-thirds are in Medicare managed care (Medicare Advantage or other private plans) or accountable care organization (ACO) models.
• The Medicare Shared Savings Program—a permanent ACO model established through the Affordable Care Act of 2010—accounts for most of the beneficiaries assigned to ACO or ACO-like payment models.
• Only 33 percent of Medicare beneficiaries with both Part A and Part B coverage are now in traditional fee-for-service (FFS) Medicare—a share that has declined in recent years.
• Even among the one-third of beneficiaries in traditional FFS, some beneficiaries may be assigned to other alternative payments models such as the Bundled Payments for Care Improvement Advanced model or the Comprehensive Primary Care Plus model.
Traditional FFS33%
MSSP ACO20%
Medicare managed care
44%
Other ACOs and ACO-like models 4%
46 Alternative payment models
Chart 5-2. The number of beneficiaries assigned to MSSP ACOs grew rapidly through 2018 but more moderately since then
Note: MSSP (Medicare Shared Savings Program), ACO (accountable care organization). Numbers are as of January in each
year. In 2019, MSSP ACOs were allowed to join the program in July 2019. Those ACOs and the beneficiaries assigned to them were not in the program as of January 2019 and are therefore not included in the 2019 counts on this chart. As of July 2019, there were 518 MSSP ACOs and 10.9 million beneficiaries assigned to them.
Source: CMS Shared Savings Program January 2020 Fast Facts.
• The number of beneficiaries assigned to MSSP ACOs grew rapidly through 2018 but has grown at a more moderate pace in recent years.
• The number of ACOs peaked in 2018 and then declined between 2018 and 2020.
• From 2018 to 2020, the simultaneous decline in MSSP ACOs but increase in assigned beneficiaries reflects larger assignment per ACO.
• CMS finalized changes to the MSSP program at the end of 2018 that included (1) requiring ACOs to transition toward greater levels of risk and (2) using regional spending as a component of all ACO benchmarks (the spending level used to measure an ACO’s financial performance). These changes coincided with some ACOs dropping out of the program and fewer new ACOs joining the program.
220
338
404433
480
561
487517
3.2
4.9
7.37.7
9.0
10.5 10.4
11.2
0
100
200
300
400
500
600
0
2
4
6
8
10
12
14
2013 2014 2015 2016 2017 2018 2019 2020
Nu
mb
er
of
MS
SP
AC
Os
MS
SP
-assig
ned
ben
efi
cia
ries (
mil
lio
ns)
A Data Book: Health care spending and the Medicare program, July 2020 47
Chart 5-3. Bundled Payments for Care Improvement (BPCI) Advanced is Medicare’s largest episode-based payment model, 2020
Sources: CMS’s Oncology Care Model website (https://innovation.cms.gov/innovation-models/oncology-care); information on the latest number of Comprehensive Care for Joint Replacement participants: Personal communication from CMS staff, May 1, 2020; information on BPCI Advanced participants: CMS's Where Innovation Is Happening website
(https://innovation.cms.gov/innovation-models/map#model=bpci-advanced).
• Medicare fee-for-service (FFS) providers can participate in episode-based payment models.
• Episode-based payment models give health care providers a spending target for most types of care provided during a clinical episode (e.g., 6 months of chemotherapy, an inpatient admission or outpatient procedure plus most other care provided in the subsequent 90 days). If total spending is less than the target, Medicare pays providers a bonus; if total spending is more than the target, Medicare recoups money from providers.
• Within FFS Medicare, the episode-based payment model with broadest participation (1,295 acute care hospitals and physician group practices participating) is the BPCI Advanced model.
• BPCI Advanced allows hospitals and practices to participate in dozens of clinical episodes, most of which are for inpatient admissions (as opposed to outpatient procedures). The most commonly pursued types of clinical episodes in BPCI Advanced are shown in Chart 5-4.
• About two-thirds of BPCI Advanced participants accept episode-based payments for fewer than six types of clinical episodes at a time. Twenty-two percent accept episode-based payments for only one type of clinical episode (data not shown).
BPCI Advanced
Comprehensive Care for Joint Replacement
Oncology Care Model
1,295 hospitals and practices
175 practices
369 mandatory + 103 voluntary = 472 hospitals
Number of participating health care organizations
48 Alternative payment models
Chart 5-4. Share of BPCI Advanced participants accepting financial responsibility for particular types of clinical episodes, 2020
Note: BPCI (Bundled Payments for Care Improvement). BPCI Advanced participants can accept episode-based payments for multiple types of clinical episodes. The denominator is 1,295 BPCI Advanced participants in 2020.
Source: List of clinical episodes each BPCI Advanced participant agreed to take financial responsibility for in Model Year 3 (2020) downloaded from CMS’s BPCI Advanced webpage (https://innovation.cms.gov/innovation-models/bpci-advanced).
10%
10%
11%
11%
12%
15%
18%
18%
19%
22%
22%
26%
27%
29%
30%
30%
33%
33%
37%
37%
39%
39%
41%
42%
48%
48%
50%
51%
52%
53%
69%
Fractures of the femur and hip or pelvis
Endovascular cardiac valve replacement
Inflammatory bowel disease
Bariatric surgery
Lower extremity and humerus procedure except hip, foot, femur
Major bowel procedure
Disorders of liver except malignancy, cirrhosis, or alcoholic hepatitis
Double joint replacement of the lower extremity
Cardiac valve
Hip and femur procedures except major joint
Coronary artery bypass graft
Pacemaker
Cellulitis
Seizures
Gastrointestinal obstruction
Major joint replacement of the upper extremity
Spinal fusion
Gastrointestinal hemorrhage
Urinary tract infection
Cardiac defibrillator
Stroke
Major joint replacement of the lower extremity
Congestive heart failure
Renal failure
Simple pneumonia and respiratory infections
Acute myocardial infarction
Chronic obstructive pulmonary disease, bronchitis, asthma
Back and neck except spinal fusion
Percutaneous coronary intervention
Cardiac arrhythmia
Sepsis
A Data Book: Health care spending and the Medicare program, July 2020 49
Chart 5-5. 2,825 practices are testing the Comprehensive Primary Care Plus (CPC+) model, 2020
Note: Comprehensive Primary Care Plus (CPC+) is an advanced alternative payment model that CMS began testing in 2017 in
some regions and in 2018 in others. CPC+ is a multipayer model, with some Medicaid and private insurers voluntarily paying
similar fees for their enrollees. Alaska (not shown) was not selected as a region eligible to participate in the CPC+ model. Source: CMS's list of CPC+ practices (https://data.cms.gov/Special-Programs-Initiatives-Speed-Adoption-of-Bes/Comprehensive-
Primary-Care-Plus/eevd-hiep).
• CMS’s CPC+ is an advanced alternative payment model that aims to strengthen primary care by
providing additional, up-front payments to participating providers of primary care services. These
payments are intended to support enhanced, coordinated care management and assist with care
delivery transformation.
• Participating practices receive a risk-adjusted per beneficiary per month care management fee, in
addition to standard fee-for-service (FFS) payments. Practices can also opt to shift some of their FFS
revenue into prospective payments received quarterly.
• CPC+ practices can earn performance bonuses unless they also participate in a Medicare Shared Savings Program (MSSP) accountable care organization (since bonuses are already available through the MSSP). About half the CPC+ practices also participate in the MSSP.
553
405 396
196 190 170 158 150 13498 81 70 55 48 33 31 30 27
OH
& N
. K
entu
cky NJ
MI
Phila
delp
hia
, P
A
CO
AR
OK
OR
Alb
an
y &
N.
Hudson,
NY
Kansas C
ity, K
S/M
O HI
Buffalo
, N
Y
TN
MT RI
NE
LA
ND
Nu
mb
er
of
pra
cti
ces
50 Alternative payment models
Chart 5-6. About 95 percent of the clinicians who qualified for a 5 percent A–APM bonus in 2020 were in ACO models
Note: A–APM (advanced alternative payment model), ACO (accountable care organization), ESRD (end-stage renal disease).
Clinicians’ 2018 A–APM participation determines their 2020 bonuses. To qualify for the A–APM bonus, clinicians had to receive 25 percent of their professional services payments or provide 20 percent of their patients with professional services through an A–APM in 2018. The A–APM bonus is equal to 5 percent of a clinician’s professional services payments from
Medicare (not including cost sharing paid by beneficiaries). In addition to the A–APMs shown above, clinicians had the option of qualifying for the A–APM bonus through participation in the Oncology Care Model (under which no clinicians qualified) or the Bundled Payments for Care Improvement Advanced model (under which one clinician qualified). For the payment models
shown, only those model tracks that require clinicians to take on some financial risk qualify as A–APMs (e.g., physicians participating in Track 1 of the Medicare Shared Savings Program did not qualify for A–APM bonuses because Track 1 involved no financial risk for participants). Percentages do not total 100 because of rounding.
Source: CMS data on clinicians who qualified for the 5 percent bonus in 2020 based on clin icians’ 2018 model participation.
• The payment models that CMS has designated as A–APMs place health care providers at some financial risk for Medicare spending while expecting them to meet quality goals for a defined patient population. Clinicians who participate in A–APMs qualify for bonuses equal to 5 percent of their professional services payments from Medicare. These bonus payments are available from 2019 to 2024.
• In 2020, about 183,000 clinicians nationwide qualified for the A–APM bonus (based on 2018 A–APM participation). About 95 percent of these clinicians participated in ACOs, which give clinicians an opportunity to earn shared savings payments from Medicare if they lower health care spending while meeting care quality standards.
• Among physicians who qualified for an A–APM bonus in 2020, 62 percent were specialists and 38 percent were primary care physicians.
Medicare Shared Savings Program 64.7%
Next Generation ACO 29.9%
Comprehensive ESRD Care 0.9%
Comprehensive Primary Care Plus 4.0%
Maryland’s Care Redesign Program and Primary Care Program
0.4%
ACO models (95.4%)
6
A Data Book: Health care spending and the Medicare program, July 2020 53
Chart 6-1. Urban IPPS hospitals accounted for under half of the 4,700 short-term acute care hospitals but about 80 percent of all-payer and Medicare FFS inpatient stays in 2018
Hospitals Inpatient stays
All payer Medicare FFS
Number Share of Number Share of Number Share of Hospital group (in thousands) total (in millions) total (in millions) total All short-term acute 4.7 100 31.8 100 9.5 100 IPPS 3.2 68 29.9 94 9.0 94 Metropolitan (urban) 2.1 45 26.3 83 7.6 80 Rural micropolitan 0.3 7 1.6 5 0.5 6 Other rural 0.7 16 2.1 7 0.9 9 For-profit 0.8 17 5 16 1.5 16 Nonprofit 1.9 40 20.9 66 6.4 67 Government 0.5 10 4.0 13 1.1 11 DSH 2.7 57 27.8 87 8.2 86 Non-DSH 0.5 11 2.2 7 0.8 8 Teaching 1.2 25 19.6 62 5.5 58 Nonteaching 2.0 43 10.4 33 3.5 37 Sole community 0.4 10 1.9 6 0.8 8 Medicare dependent 0.2 4 0.3 1 0.1 1 Neither 2.6 55 27.7 87 8.0 85 Critical access 1.3 29 0.6 2 0.3 3 Maryland <0.1 1 0.6 2 0.2 2 Children’s <0.1 2 0.6 2 <0.1 <1 Cancer <0.1 <1 0.1 0 <0.1 <1 Note: IPPS (inpatient prospective payment system), FFS (fee-for-service), DSH (disproportionate share hospital). Data are for
short-term acute care hospitals in the U.S. (excluding territories) that had a cost report with a midpoint in fiscal year 2018.
“Number of hospitals” is the number of Medicare provider numbers; a single provider number can represent multiple hospital locations. Metropolitan (urban) counties contain an urban cluster of 50,000 or more people, and rural micropolitan counties contain a cluster of 10,000 to 50,000 people. Components may not sum to totals due to rounding.
Source: MedPAC analysis of hospital cost report data from CMS.
• In 2018, there were approximately 4,700 short-term acute care hospitals participating in the Medicare program, including 3,220 paid under the inpatient prospective payment system and 1,350 small, rural hospitals designated as critical access hospitals.
• Metropolitan IPPS hospitals accounted for 45 percent of short-term acute care hospitals but 83 percent of the 31.8 million all-payer inpatient stays and 80 percent of the 9.5 million Medicare FFS inpatient stays.
54 Acute inpatient services
Chart 6-2. The number of general, short-term acute care hospitals that ceased inpatient services substantially increased in 2019
Note: “Closure” refers to a hospital location that ceased inpatient services, while “opening” refers to a new location for inpatient
services. The chart does not include the relocation of inpatient services from one hospital to another under common ownership within 10 miles, nor does it include hospitals that both opened and closed within a 5-year time period. Data are for general short-term acute care hospitals in the U.S. paid under the inpatient prospective payment system, designated
as critical access hospitals, or covered under the Maryland state waiver. Metropolitan (urban) counties contain an urban cluster of 50,000 or more people, and rural micropolitan counties contain a cluster of 10,000 to 50,000 people. The counts in this chart differ from those previously published for several reasons, including that this chart counts closures and openings based on fiscal year, uses an updated methodology, and is updated to remove hospitals previously counted as
closures but that have since reopened.
Source: MedPAC analysis of the CMS Provider of Services file, census data on metropolitan and micropolitan areas, internet searches, and personal communication with the Department of Health and Human Services Office of Rural Health Policy.
• While hospital closures are still relatively rare events, there was a substantial increase in the number of hospitals that ceased inpatient services in fiscal year 2019, without a corresponding increase in openings.
• In 2019, 46 general short-term acute care hospitals participating in the Medicare program closed, and 11 hospitals opened. Among the 46 closures, 26 were in metropolitan counties, 7 were in rural micropolitan counties, and 13 were in other rural counties.
• The hospitals that closed in 2019 tended to be small (30 had 100 or fewer beds), had low inpatient occupancy rates (approximately 25 percent, on average), and had poor profitability (all-payer margin of –16 percent, on average, in the year before closure) (data not shown).
• Nearly all of the hospital openings from 2015 to 2019 were in metropolitan counties.
21
9 914
26
3
4 3
4
74
84
5
13
28
2116
23
46
0
10
20
30
40
50
60
2015 2016 2017 2018 2019
Nu
mb
er
of
ho
sp
itals
Closures
116 10 12 10
128 10 12 11
0
10
20
30
2015 2016 2017 2018 2019
Nu
mb
er
of
ho
sp
itals
Fiscal year
Metropolitan Rural micropolitan Other rural
Openings
A Data Book: Health care spending and the Medicare program, July 2020 55
Chart 6-3. Aggregate occupancy rate at short-term acute care hospitals has increased slightly, but remained much lower at rural hospitals, 2014–2018
Note: Hospital occupancy rates are defined as total bed days (including swing bed days) and observation bed days used, minus
nursery bed days used, divided by total bed days available. Data are for short-term acute care hospitals in the U.S. (excluding territories) that had a cost report with a midpoint in fiscal year 2018. Metropolitan (urban) counties contain an urban cluster of 50,000 or more people, and rural micropolitan counties contain a cluster of 10,000 to 50,000 people.
Source: MedPAC analysis of hospital cost report data from CMS.
• The aggregate occupancy rate at short-term acute care hospitals increased slightly between 2014 and 2018, from 61 percent to 64 percent.
• Occupancy rates are generally higher for metropolitan hospitals than rural micropolitan or other rural hospitals. However, occupancy rates rose the fastest for rural micropolitan hospitals during this five-year period, with an average annual growth rate of 1.9 percent.
• Increasing occupancy in metropolitan and rural micropolitan areas in part reflects increasing volumes of inpatient days at these hospitals. In contrast, the total number of inpatient days in other rural areas declined, but the aggregate occupancy rate remained relatively steady because available inpatient bed days also declined due to closures and other reductions in beds (data not shown).
64%66% 66% 67% 67%
61% 62% 63% 63% 64%
50%52% 53% 54% 55%
41% 41% 40% 40% 41%
0%
10%
20%
30%
40%
50%
60%
70%
80%
2014 2015 2016 2017 2018
Ag
gre
gate
occu
pan
cy r
ate
Fiscal year
Metropolitan
All
Rural micropolitan
Other rural
56 Acute inpatient services
Chart 6-4. Divergent trends in all-payer hospital outpatient visits and inpatient stays per capita narrowed in 2017 and 2018
Note: “Outpatient visits” includes all clinic visits, referred visits, observation services, outpatient surgeries, and emergency department visits, regardless of the number of diagnostic and/or therapeutic treatments the patient received during the
visit. Data are for community hospitals (nonfederal short-term general and specialty hospitals), estimated from those who responded to the American Hospital Association (AHA) survey. With the 2019 edition of Hospital Statistics, the AHA began using a new methodology to classify facilities as hospitals. As a result of the application of the new, broader
hospital definition, the number of community hospitals in each year from 2013 to 2017 increased by approximately 400.
Source: MedPAC analysis of Hospital Statistics data from the American Hospital Association and CMS National Health
Expenditure data.
• From 2010 to 2014, there were divergent trends in all-payer hospital outpatient visits and inpatient stays per capita, with growth in outpatient visits and declines in inpatient stays.
• Beginning in 2015, the divergent trends in all-payer outpatient and inpatient growth rates started to narrow, as inpatient stays per capita held relatively steady.
• Starting in 2017, the trends in all-payer outpatient visits and inpatient stays per capita were similar: both grew slightly in 2017 (0.6 percent and 0.2 percent, respectively) and decreased slightly in 2018 (–0.6 percent and –0.8 percent).
-1.9%-1.5%
-1.9% -1.7%
-2.3%
0.2% -0.1%0.2%
-0.8%
0.7%
0.0%
2.1%
0.7%
1.6%
3.6%
2.8%
0.6%
-0.6%
-5%
-4%
-3%
-2%
-1%
0%
1%
2%
3%
4%
5%
2010 2011 2012 2013 2014 2015 2016 2017 2018
An
nu
al p
erc
en
tag
e c
han
ge
Fiscal year
Oupatient visits per capita
Inpatient staysper capita
A Data Book: Health care spending and the Medicare program, July 2020 57
Chart 6-5. Hospital employment growth from 2015 to 2019 was driven by an increase in health care practitioners
Note: NP (nurse practitioner), PA (physician assistant). Data are for general medical and surgical hospitals. The percentages
shown at the top of each category are the cumulative percentage change from 2015 to 2019. Components of other health
care practitioners in the bottom chart do not sum to all other health care practitioners in the top chart because only some subsets of practitioner occupations are shown.
Source: MedPAC analysis of Bureau of Labor Statistics, Current Employment Statistics data.
• The Bureau of Labor Statistics survey of employers indicates that general medical and surgical hospitals employed 5.6 million individuals in 2019. Of these, approximately 1.7 million (31 percent) were registered nurses and 1.4 million (26 percent) were other health care practitioners. The remaining 44 percent of hospital employees were in nonpractitioner occupational categories.
• From 2015 to 2019, the number of registered nurses increased 7.9 percent, similar to the rate of all other health care practitioners. However, within the group of other health care practitioners, several occupations grew rapidly. For example, while still a small share of all health care practitioners, the number of nurse practitioners and physician assistants employed by hospitals increased 35.9 percent.
• From 2015 to 2019, the number of hospital staff in several nonpractitioner occupations increased as well. For example, health care support staff (such as nursing assistants and orderlies) increased 8.3 percent.
1,5871,339
643 695
1,011
1,7131,438
696 657
1,078
0
500
1,000
1,500
2,000
2,500
3,000
Registered nurses Other health carepractitioners
Health caresupport
Office andadministrative
support
All other
Nu
mb
er
em
plo
yed
(t
ho
usan
ds)
7.9% 7.4% 8.3% –5.5% 6.6%
147
57
202223
160172
77
222245
170
0
100
200
300
400
500
Physicians andsurgeons
NPs and PAs Therapists Diagnostictechnicians
Health supporttechnicians
Nu
mb
er
em
plo
yed
(t
ho
usan
ds)
2015 2016 2017 2018 2019
16.6% 35.9% 9.4% 10.0% 6.1%
Selected occupations
58 Acute inpatient services
Chart 6-6. Short-term acute care hospitals’ all-payer financial performance remained strong, 2014–2018
Note: EBITDA (earnings before interest, taxes, depreciation, and amortization). Data are for short-term acute care hospitals in
the U.S. covered under the inpatient prospective payment system (excluding territories and those that report all-inclusive rates), that had complete cost reports and non-outlier cost per stay data. Aggregate margin is calculated as revenue minus costs, divided by revenue. “Total all-payer margin” includes all patient care services funded by all payers plus
nonpatient revenue such as investment income.
Source: MedPAC analysis of hospital cost report data from CMS.
• Hospitals’ aggregate margin for all payers—Medicare, Medicaid, other government, and private payers—reflects the relationship of all hospital revenues to all hospital costs, including inpatient, outpatient, post-acute care, and nonpatient services.
• In 2018, hospitals’ aggregate total all-payer margin (which includes investment income) was 6.8 percent, a slight decrease from the all-time high of 7.1 percent in 2017 and 2014.
• Other measures of all-payer profitability also remained strong. Hospitals’ cash flow—as measured by EBITDA—has remained steady and strong for the decade, with an aggregate EBITDA margin between 10 percent and 11 percent. Hospitals’ operating margin also remained steady and strong.
10.8 10.610.2 10.3 10.6
7.1 6.96.5
7.1 6.8
6.06.4
5.9 5.9 6.1
0
2
4
6
8
10
12
14
2014 2015 2016 2017 2018
Ag
gre
gate
marg
in (
perc
en
t)
Fiscal year
EBITDA margin
Total all-payer margin
Operating margin
A Data Book: Health care spending and the Medicare program, July 2020 59
Chart 6-7. Urban IPPS hospitals continue to have a higher aggregate total all-payer margin than rural IPPS or critical access hospitals, 2014–2018
Note: IPPS (inpatient prospective payment system). Data are for short-term acute care hospitals in the U.S. paid under the inpatient prospective payment system (excluding territories and those that report all-inclusive rates) or are designated as critical access hospitals that had complete cost reports and non-outlier cost per stay data. Aggregate margin is calculated
as revenue minus costs, divided by revenue. “Total all-payer margin” includes all patient care services funded by all payers, plus nonpatient revenue such as investment income. Metropolitan (urban) counties contain an urban cluster of 50,000 or more people; all other counties are classified as rural.
Source: MedPAC analysis of hospital cost report data from CMS.
• Metropolitan (urban) IPPS hospitals continue to have a higher aggregate total all-payer margin than rural IPPS hospitals or critical access hospitals.
• From 2017 to 2018, the aggregate total all-payer margin for metropolitan IPPS hospitals decreased slightly from 7.3 to 6.9 percent, while the margin for rural IPPS hospitals decreased from a relative high of 5.7 percent (the highest margin since 2007 (data not shown)) to 4.4 percent.
• From 2017 to 2018, the aggregate total all-payer margin for critical access hospitals also decreased, from 3.6 percent to 2.7 percent (the lowest margin since 2010 (data not shown)).
7.37.0
6.6
7.36.9
4.8 4.84.6
5.7
4.4
3.4
4.4
3.6 3.6
2.7
0
1
2
3
4
5
6
7
8
9
10
2014 2015 2016 2017 2018
Ag
gre
gate
to
tal all
-payer
marg
in (
perc
en
t)
Fiscal year
Metropolitan IPPS
Rural IPPS
Critical access hospitals
60 Acute inpatient services
Chart 6-8. Major teaching hospitals continue to have a lower aggregate total all-payer margin than nonteaching and other teaching hospitals, 2014–2018
Note: Data are for short-term acute care hospitals in the U.S. covered under the inpatient prospective payment system (excluding territories and those that report all-inclusive rates) that had complete cost reports and non-outlier cost per stay data. Aggregate margin is calculated as revenue minus costs, divided by revenue. Total margin includes all patient care
services funded by all payers, plus nonpatient revenue such as investment income. “Major teaching” hospitals are defined by a ratio of interns and residents to beds of 0.25 or greater, while “other teaching” hospitals have a ratio of greater than 0 and less than 0.25.
Source: MedPAC analysis of hospital cost report data from CMS.
• The aggregate total all-payer margin for major teaching hospitals continues to be lower than that for other teaching and nonteaching hospitals. (In contrast, major teaching hospitals have a higher Medicare overall margin than other hospitals, see Chart 6-15.)
• From 2017 to 2018, the aggregate total all-payer margin for major teaching hospitals fell from 5.5 to 4.9 percent. Over this same period, the aggregate total all payer-margin decreased slightly less for nonteaching hospitals (from 7.3 percent to 7.0 percent) and for other teaching hospitals (from 8.6 percent to 8.5 percent).
• While the aggregate total all-payer margin for major teaching, other teaching, and nonteaching hospitals each decreased from 2017 to 2018, they each remained well above levels reported from 1997 to 2012 (data not shown).
8.8
7.87.4
8.6 8.5
7.5 7.37.0
7.37.0
4.95.3
4.9
5.5
4.9
0
1
2
3
4
5
6
7
8
9
10
2014 2015 2016 2017 2018
Ag
gre
gate
to
tal a
ll-p
ayer
marg
in (
perc
en
t)
Fiscal year
Other teaching
Nonteaching
Major teaching
A Data Book: Health care spending and the Medicare program, July 2020 61
Chart 6-9. Inpatient services are the largest component of Medicare FFS revenue for IPPS hospitals, while outpatient services are the largest and fastest growing for CAHs, 2014–2018
Note: FFS (fee-for-service), IPPS (inpatient prospective payment system), CAH (critical access hospital), UC (uncompensated
care), DGME (direct graduate medical education), HIT (health information technology). Data are for the approximately 3,200 IPPS hospitals and 1,350 CAHs in the U.S. (excluding territories) that had a cost report with a midpoint in fiscal year 2018. Revenue includes payments from the Medicare FFS program (reflective of bad debt) and beneficiary cost sharing
for services provided to Medicare FFS beneficiaries and the FFS share of supplemental payments. It does not include Medicare program payments for Medicare Advantage beneficiaries. Post-acute care includes swing bed, rehabilitation, and skilled nursing services. Components may not sum to totals due to rounding.
Source: MedPAC analysis of hospital cost report data from CMS.
• The approximately 3,200 general short-term acute care hospitals paid under the IPPS received $189 billion in Medicare FFS revenue in 2018, including $112 billion for general inpatient services and $60 billion for outpatient services. From 2014 to 2018, IPPS hospitals’ Medicare FFS inpatient revenue increased at an average annual rate of 1.4 percent, while outpatient revenue increased 6.2 percent. These increases were driven by increases in payments per service (data not shown).
• The approximately 1,350 critical access hospitals (CAHs) received $11 billion in Medicare FFS revenue in 2018, including $2.8 billion for general inpatient services, $6.2 billion for outpatient services, and $2.1 billion in post-acute care services (mainly provided in swing beds). From 2014 to 2018, CAHs’ Medicare FFS inpatient revenue increased at an average annual rate of 0.9 percent, while outpatient revenue increased 6.7 percent, and post-acute care revenue increased 4.3 percent. These increases were driven by increases in payments per service (data not shown).
106 107 110 112 112
47 50 53 56 60
173 176 179 185 189
0
50
100
150
200
250
2014 2015 2016 2017 2018
Bil
lio
ns o
f d
oll
ars
IPPS
2.7 2.7 2.7 2.8 2.8
4.7 5.0 5.4 5.8 6.2
1.8 1.9 2.0 2.0 2.19.5 9.8 10.1 10.6 11.0
0
5
10
15
2014 2015 2016 2017 2018
Bil
lio
ns o
f d
oll
ars
Fiscal year
General inpatient services Outpatient services
Post-acute care and psychiatric services Supplemental payments (UC, DGME, HIT)
CAH
62 Acute inpatient services
Chart 6-10. Base PPS payments represented about 84 percent of IPPS hospitals’ overall Medicare FFS revenue, 2018
Share of overall Medicare FFS revenue Base Low High Rural Cost UC Bad debt PPS (and income, cost and/or pass and not Hospital group short staya) teachingb outliers isolatedc through DGME reimbursed All inpatient PPS 83.7% 5.6% 3.1% 2.0% 1.4% 5.1% –0.6% Metropolitan 83.4 6.2 3.3 1.3 1.5 5.4 –0.5 Rural micropolitan 88.8 2.3 2.4 3.4 0.6 3.0 –0.6 Other rural 84.0 2.0 1.4 7.9 0.8 3.1 –0.8
For profit 88.5 4.2 2.3 1.5 0.5 4.6 –0.8 Nonprofit 84.3 5.5 3.0 2.1 1.2 4.7 –0.5 Government 76.0 7.3 4.1 2.1 3.1 7.8 –0.6
DSH 82.9 5.9 3.1 2.0 1.4 5.5 –0.6 Non-DSH 92.9 1.6 2.6 1.4 0.9 0.5 –0.3
Teaching 80.3 7.6 3.5 1.6 1.8 6.2 –0.5 Nonteaching 90.4 1.6 2.2 2.7 0.5 3.0 –0.6
Sole community 80.2 2.5 2.2 11.5 0.3 2.8 –0.6 Medicare dependent 82.7 1.2 0.9 11.3 0.1 2.4 –0.9 Neither 84.1 6.0 3.2 0.9 1.5 5.4 –0.6
Critical access 1.6 0.0 0.0 0.1 99.3 0.0 –1.1 Note: PPS (prospective payment system), IPPS (inpatient prospective payment system), FFS (fee-for-service), DSH
(disproportionate share hospital), UC (uncompensated care), DGME (direct graduate medical education). “Overall Medicare FFS revenue” includes payments from the Medicare FFS program (reflective of bad debt) and beneficiary cost sharing for services provided to Medicare FFS beneficiaries and their share of supplemental payments, across hospital
service lines (inpatient, outpatient, and swing bed, rehabilitation, skilled nursing, and psychiatric services). Metropolitan (urban) counties contain an urban cluster of 50,000 or more people, and rural micropolitan counties contain a cluster of 10,000 to 50,000 people. Components may not sum to totals because other types of payments, such as quality,
demonstration, and reconciliation payments, are not included in the table. a"Short stay" includes adjustments for short-stay transfers in the inpatient PPS and short-stay outlier payments in the
inpatient rehabilitation and psychiatric PPSs. b"Low income, teaching" includes the disproportionate share and indirect medical education payments in the inpatient
PPS and the low-income adjustment in the inpatient rehabilitation PPS. c“Rural and/or isolated” includes payments above federal inpatient PPS rates from sole community hospital– or Medicare-
dependent hospital–specific rates, the low-volume adjustment, and the rural adjustments to the rehabilitation and psychiatric PPSs.
Source: MedPAC analysis of hospital cost report data from CMS.
• Base PPS payments accounted for about 84 percent of IPPS hospitals’ aggregate overall Medicare FFS revenue, while PPS low-income and teaching adjustments, outlier payments, rural and/or isolated payments, cost-based pass-through amounts, supplemental payments, and bad debt accounted for the remaining 16 percent. However, the share of Medicare FFS revenue from different payment types varied substantially across different groups of IPPS hospitals.
• Cost-based reimbursement for CAHs results in payments significantly above what CAHs would be paid under the hospital PPSs, and it results in higher beneficiary cost sharing. (CAHs can have some PPS payments if they operate distinct-part rehabilitation or psychiatric hospitals that are reimbursed for those services under the respective PPSs.)
A Data Book: Health care spending and the Medicare program, July 2020 63
Chart 6-11. Since implementation of uncompensated care payments in 2014, DSH payments have slowly increased, while uncompensated care payments have varied
Note: DSH (disproportionate share). Payments represent CMS’s estimated operating DSH payments and uncompensated care
payments, before sequestration. Chart does not include capital DSH payments.
Source: CMS IPPS final rules.
• In each of 2012 and 2013, IPPS hospitals received approximately $12 billion in aggregate operating DSH payments. The traditional DSH payment formula is based on hospitals’ share of Medicaid patients and Medicare patients with Supplemental Security Income (SSI) and therefore results in increased DSH payments as Medicaid expands.
• Beginning in 2014, a policy change reduced operating DSH payments but added uncompensated care payments. Specifically, beginning in 2014, IPPS hospitals’ operating DSH payments were calculated as 25 percent of the DSH payment the hospital would have received under the traditional DSH formula in effect before 2014. At the same time, a fixed pool of uncompensated care payments was created, set each year at 75 percent of the estimated aggregate operating DSH payments IPPS hospitals would have received under the traditional formula, less a percentage reduction in the uncompensated care pool that is proportional to the decline in the rate of uninsured since 2013. This uncompensated care pool is distributed to DSH-eligible IPPS hospitals based on each hospital’s share of aggregate uncompensated care.
• Since the policy change, aggregate operating DSH payments have been slowly increasing, from $3.2 billion in 2014 to $4.1 billion in 2020. However, the amount of uncompensated care payments has varied, consistent with trends in the share of the population with Medicaid and without health insurance.
3.2 3.3 3.4 3.6 3.9 4.1 4.1
9.07.6
6.4 6.06.8
8.3 8.4
11.7 12.1 12.2
10.9
9.8 9.6
10.7
12.4 12.5
0
2
4
6
8
10
12
14
16
2012 2013 2014 2015 2016 2017 2018 2019 2020
Bil
lio
ns o
f d
oll
ars
Fiscal year
DSH payments Uncompensated care payments
64 Acute inpatient services
Chart 6-12. Financial pressure led to lower hospital costs per discharge, 2015–2017
Level of financial pressure, 2015–2017
High pressure Low pressure (non-Medicare Medium (non-Medicare margin ≤ 1%) pressure margin > 5%)
Number of hospitals 666 337 1,729
Financial characteristics, 2018 (medians)
Non-Medicare margin (private, Medicaid, uninsured) –4% 3% 14%
Standardized cost per discharge (as a share of the national median) For-profit and nonprofit hospitals 0.96 0.97 1.02 Nonprofit hospitals 0.97 0.99 1.04 For-profit hospitals 0.90 0.92 0.94 Annual growth in cost per discharge, 2016–2018 2% 2% 2% Overall 2018 Medicare margin (medians) –1% –4% –10%
Patient characteristics (medians) Total hospital discharges in 2018 3,347 6,483 7,872 Medicare share of inpatient days 39% 37% 37% Medicaid share of inpatient days 8% 7% 6% Medicare case-mix index 1.43 1.53 1.65
Note: Standardized costs are adjusted for hospital case mix, wage index, outliers, transfer cases, interest expense, and the
effects of teaching and low-income Medicare patients on hospital costs. The sample includes short-term acute care hospitals paid under the inpatient prospective payment system that had complete cost reports on file with CMS by October 2019. “High-pressure” hospitals are defined as those with a median non-Medicare profit margin of 1 percent or
less from 2015 to 2017 and a net worth (assets minus liabilities) that grew by less than 1 percent per year over that period if the hospital’s Medicare profits had been zero. “Low-pressure” hospitals are defined as those with a median non-Medicare profit margin greater than 5 percent from 2015 to 2017 and a net worth that grew by more than 1 percent per year over that period if the hospital’s Medicare profits had been zero. “Medium-pressure” hospitals are those that fit into
neither the high- nor the low-pressure categories.
Source: MedPAC analysis of hospital cost report data and claims files from CMS.
• Hospitals under higher financial pressure had 4 percent lower standardized costs per discharge than the national median. For-profit hospitals tended to constrain their costs more than nonprofit hospitals. The median for-profit hospital had costs that were 6 percent below the average even when they were not under financial pressure.
• Hospitals with lower volume, lower case mix, and higher Medicaid and Medicare shares of discharges are more likely to be under financial pressure.
• One limitation of this analysis is that it measures only hospital inpatient costs. To the extent that hospitals with strong profit margins direct their resources toward non-inpatient expenditures (such as the purchase or subsidization of physician practices), those costs would not be included in our standardized costs per discharge.
A Data Book: Health care spending and the Medicare program, July 2020 65
Chart 6-13. Aggregate overall Medicare margin for short-term acute care hospitals increased slightly from 2017 to 2018
Note: Data are for short-term acute care hospitals in the U.S. covered under the inpatient prospective payment system (IPPS)
(excluding territories and those that report all-inclusive rates), that had complete cost reports and non-outlier cost per stay data. Aggregate margin is calculated as revenue minus costs, divided by revenue. Margins are based on Medicare-allowable costs. The overall Medicare margin includes the costs and payments of acute inpatient, outpatient, inpatient
psychiatric, rehabilitation, skilled nursing facility, and home health services, as well as graduate medical education, bad debts, health information technology, and uncompensated care payments.
Source: MedPAC analysis of hospital cost report data from CMS.
• The aggregate overall Medicare margin incorporates payments and costs for acute inpatient, outpatient, skilled nursing, home health care, and inpatient psychiatric and rehabilitative services, as well as direct graduate medical education, bad debts, Medicare
payments for health information technology, and—starting in 2014—uncompensated care
payments.
• From 2017 to 2018, the aggregate overall Medicare margin for IPPS hospitals increased from –9.9 percent to –9.3 percent. However, the margin remains well below levels from 2010 to 2013.
• The range of overall Medicare margins at individual IPPS hospitals varied substantially. For example, in 2018, 25 percent of hospitals had an overall Medicare margin of 1.8 percent or higher, and another 25 percent had a margin of –19.2 percent or lower (data not shown).
-4.9
-5.7-5.4
-5.0-5.6
-7.6
-9.7 -9.9-9.3
-12
-10
-8
-6
-4
-2
0
2010 2011 2012 2013 2014 2015 2016 2017 2018
Ag
gre
gate
overa
ll M
ed
icare
marg
in
(perc
en
t)
Fiscal year
66 Acute inpatient services
Chart 6-14. Rural short-term acute care hospitals continue to have a higher aggregate overall Medicare margin than urban hospitals
Note: CAH (critical access hospital), IPPS (inpatient prospective payment system). Data are for short-term acute care hospitals in the U.S. covered under the inpatient prospective payment system (excluding territories and those that report all-
inclusive rates), that had complete cost reports and non-outlier cost per stay data. Aggregate margin is calculated as revenue minus costs, divided by revenue. Margins are based on Medicare-allowable costs. The overall Medicare margin includes the costs and payments of acute inpatient, outpatient, inpatient psychiatric, rehabilitation, skilled nursing facility,
and home health services, as well as graduate medical education, bad debts, health information technology, and uncompensated care payments. Metropolitan (urban) counties contain an urban cluster of 50,000 or more people; all other counties and all CAHs are classified as rural.
Source: MedPAC analysis of hospital cost report data from CMS.
• Since 2005, the aggregate overall Medicare margin for short-term acute care hospitals in rural counties has exceeded that for hospitals in metropolitan counties (not all data shown). The higher rural margins reflect special rural add-on payments (see Chart 6-10), including the introduction of low-volume add-on payments in 2013.
• In 2018, the difference between the aggregate overall Medicare margin at metropolitan and rural hospitals was 3.0 percentage points when compared with rural IPPS hospitals and 4.7 percentage points when compared with rural IPPS and critical access hospitals.
-1.9
-3.2
-5.4-5.9
-4.9
-3.5
-4.9
-7.5
-8.2
-6.6
-5.8
-7.9
-9.9 -10.0-9.6
-12
-10
-8
-6
-4
-2
0
2014 2015 2016 2017 2018
Ag
gre
gate
overa
ll M
ed
icare
marg
in
(perc
en
t)
Fiscal year
Rural (including CAHs)
Rural IPPS (excluding CAHs)
Metropolitan IPPS
A Data Book: Health care spending and the Medicare program, July 2020 67
Chart 6-15. Teaching and disproportionate share short-term acute care hospitals continue to have higher aggregate Medicare margins than other hospitals
Note: DSH (disproportionate share). Data are for short-term acute care hospitals in the U.S. covered under the inpatient
prospective payment system (excluding territories and those that report all-inclusive rates) that had complete cost reports and non-outlier cost per stay data. Aggregate margin is calculated as revenue minus costs, divided by revenue. Medicare
margin is based on Medicare-allowable costs. The overall Medicare margin includes the costs and payments of acute inpatient, outpatient, inpatient psychiatric, rehabilitation, skilled nursing facility, and home health services, as well as graduate medical education, bad debts, health information technology, and uncompensated care payments.
Source: MedPAC analysis of hospital cost report data from CMS.
• Both teaching hospitals and those that treat a large share of low-income patients (referred to as “disproportionate share hospitals”) continue to have higher aggregate overall Medicare margins than other hospitals. Their better financial performance under Medicare is largely due to the additional payments they receive from the indirect medical education and DSH adjustments to their inpatient payments, as well as supplemental uncompensated care payments.
-3.7
-6.3
-8.5
-9.0 -9.6
-5.0-6.3
-8.6
-8.2 -7.5
-7.7
-9.9
-11.7 -12.2-10.9
-20
-15
-10
-5
0
2014 2015 2016 2017 2018
Ag
gre
gate
overa
ll M
ed
icare
m
arg
in (
perc
en
t)
Major teaching
Other teaching
Nonteaching
-2.3
-4.6
-7.2-8.1 -8.3
-6.4
-8.1
-10.0 -9.9-9.1
-13.3
-15.3 -15.7-16.4
-14.7
-20
-15
-10
-5
0
2014 2015 2016 2017 2018
Ag
gre
gate
overa
ll M
ed
icare
m
arg
in (
perc
en
t)
Fiscal year
High DSH
Moderate-to-low DSH
No DSH
68 Acute inpatient services
Chart 6-16. Decline in Medicare FFS inpatient stays per capita has slowed, with larger declines at critical access and rural hospitals
Note: FFS (fee-for-service). Data are for short-term acute care hospitals in the U.S. (exclusive of territories). “General short-term
hospital” refers to short-term acute care hospitals paid under the inpatient prospective payment system or the Maryland
state waiver. Metropolitan (urban) counties contain an urban cluster of 50,000 or more people; all other counties are classified as rural.
Source: MedPAC analysis of Medicare Provider Analysis and Review data and enrollment data.
• The number of inpatient stays per 1,000 Medicare FFS beneficiaries has decreased from 306 in 2010 to 250 in 2018, declining sharply in the early years of this period but at a slowing rate in more recent years (data not shown).
• The magnitude of the decrease in inpatient stays per capita varied across types of hospitals, with larger declines at critical access hospitals and rural hospitals.
• From 2017 to 2018, the number of inpatient stays per capita fell 2.1 percent at critical access hospitals, compared with 1.6 percent at general acute care hospitals. During the same time period, the number of inpatient stays per capita fell 2.1 percent at rural hospitals, compared with 1.5 percent at metropolitan hospitals.
-1.7%
-6.1%
-3.7%-4.4%
0.3%
-2.7%
0.3%
-1.6%-1.8%
-9.5%
-5.0%
-7.8%
1.7%
-7.4%
-1.3%-2.1%
-15%
-10%
-5%
0%
5%
2011 2012 2013 2014 2015 2016 2017 2018An
nu
al p
erc
en
t ch
an
ge in
in
pati
en
t sta
ys p
er
cap
ita
General short term Critical access
-1.3%
-5.8%
-3.4%-4.2%
0.5%
-2.4%
0.4%
-1.5%
-4.0%
-9.2%
-6.2%-7.0%
-0.6%
-5.6%
-0.8%-2.1%
-15%
-10%
-5%
0%
5%
2011 2012 2013 2014 2015 2016 2017 2018
An
nu
al p
erc
en
t ch
an
ge in
in
pati
en
t sta
ys p
er
cap
ita
Fiscal year
Metropolitan Rural
A Data Book: Health care spending and the Medicare program, July 2020 69
Chart 6-17. Circulatory system diagnoses remain the most common reason for Medicare FFS beneficiaries' inpatient stays at short-term acute care hospitals, but the share has decreased, 2010–2018
Note: FFS (fee-for-service). Data are for short-term acute care hospitals in the U.S. (exclusive of territories).
Source: MedPAC analysis of Medicare Provider Analysis and Review data.
• In 2018, four major diagnostic categories accounted for over 57 percent of all Medicare FFS
inpatient stays at short-term acute care hospitals.
• The circulatory system was the most common major diagnostic category among Medicare FFS inpatient stays; however, its share declined from about 23 percent to 20 percent from 2010 to 2018. Circulatory system diagnoses include heart failure and cardiac arrhythmia.
• From 2010 to 2018, the major diagnostic category with the largest increase was infectious and parasitic diseases, which increased from about 5 percent to nearly 11 percent of Medicare FFS inpatient stays, due to growth in the number of FFS beneficiaries hospitalized with septicemia (severe sepsis).
23.3%
20.1%
12.1%13.5%
14.6%
13.4%
5.1%
10.5%
0%
5%
10%
15%
20%
25%
30%
2010 2011 2012 2013 2014 2015 2016 2017 2018
Sh
are
of
inp
ati
en
t sta
ys (
in p
erc
en
t)
Fiscal year
Circulatory system
Musculoskeletal system and connective tissue
Respiratory system
Infectious and parasitic diseases
70 Acute inpatient services
Chart 6-18. A growing share of Medicare FFS beneficiaries' inpatient stays at short-term acute care hospitals are immediately followed by post-acute care or hospice, 2010–2018
Note: FFS (fee-for-service). Data are for short-term acute care hospitals in the U.S. (exclusive of territories). Components do not
sum to 100 percent because beneficiaries discharged to other destinations are not shown.
Source: MedPAC analysis of Medicare Provider Analysis and Review data.
• From 2010 to 2018, the share of inpatient stays in which the Medicare FFS beneficiary was discharged home under self-care consistently declined, while the shares discharged to post-acute care and hospice consistently increased.
• In conjunction with the decline in Medicare FFS inpatient stays per capita, these trends could reflect in part a shift of care for less severe conditions to outpatient settings, with the remaining inpatient stays consisting of sicker patients. The increase in the share discharged to hospice also reflects increased use of hospice care in end-of-life planning.
48.9%
48.1%
47.7%
46.5%
45.6%
45.2%
45.4%
45.0%
44.8%
40.5%
41.1%
41.4%
42.5%
43.3%
43.6%
43.6%
43.9%
44.0%
2.3%
2.5%
2.7%
2.8%
2.9%
3.0%
3.0%
3.1%
3.2%
3.4%
3.4%
3.3%
3.4%
3.3%
3.4%
3.3%
3.2%
3.2%
2010
2011
2012
2013
2014
2015
2016
2017
2018
Fis
cal year
Home Post-acute care Hospice Died
A Data Book: Health care spending and the Medicare program, July 2020 71
Chart 6-19. Share of Medicare FFS beneficiaries' short stays at short-term acute care hospitals decreased from 2010 to 2014 but then began to increase again
Note: FFS (fee-for-service). Data are for short-term acute care hospitals in the U.S. (exclusive of territories). Components may not sum to 100 percent due to rounding.
Source: MedPAC analysis of Medicare Provider Analysis and Review data.
• The share of Medicare FFS beneficiaries that were one- or two-day stays decreased from 2010 to 2014, but then began to increase again.
• As the Commission has previously noted, growth in the number of one-day stays starting in 2015 could be due to the reduced likelihood that CMS’s recovery audit contractors (RACs) would deny payment for one-day stays. In 2015, CMS ceased patient status reviews (which previously resulted in challenges to one-day stay claims). The result was that from 2014 to 2015, claims challenged by the RACs as overpayments fell by 91 percent (data not shown).
• From 2017 to 2018, the share of Medicare FFS beneficiaries’ inpatient stays that were only one day increased from 12.9 to 13.4 percent, while the share of two-day stays held steady at 18.4 percent, and stays of three or more days decreased from 68.7 to 68.2 percent.
13.7%
13.3%
12.7%
12.4%
11.6%
11.6%
12.3%
12.9%
13.4%
16.1%
16.2%
16.4%
16.4%
17.2%
17.6%
18.1%
18.4%
18.4%
70.2%
70.6%
70.9%
71.3%
71.1%
70.8%
69.7%
68.7%
68.2%
2010
2011
2012
2013
2014
2015
2016
2017
2018
Fis
cal year
1 day 2 days 3+ days
72 Acute inpatient services
Chart 6-20. Number of Medicare FFS outpatient observation visits per capita decreased slightly in 2017 and 2018, and nearly half remained longer than 24 hours
Note: FFS (fee-for-service). Observation visits are separately payable visits with a length of stay of at least eight hours. Data for
outpatient observation visits include short-term acute care hospitals in the U.S. (exclusive of territories) paid under the inpatient prospective payment system or under the Maryland state waiver. “Outpatient observation visits” per capita refers to outpatient services per Medicare FFS Part B beneficiary. Years are calendar years.
Source: MedPAC analysis of outpatient standard analytical file data from CMS.
• Hospitals use observation care to determine whether a patient should be hospitalized for inpatient care, transferred to an alternative treatment setting, or sent home. On April 1, 2002, Medicare began providing separate payments to hospitals for some observation services. Previously, the observation services were packaged into the payments for the emergency department or clinic visits that occurred with observation care.
• In 2018, Medicare FFS beneficiaries had approximately 1.5 million outpatient observation visits, equivalent to 44.9 outpatient observation visits per 1,000 Medicare FFS beneficiaries.
• From 2014 to 2016, the number of Medicare FFS outpatient observation visits per capita increased by 3.2 visits. However, this slow growth reversed starting in 2016, with a decrease of 0.5 visits from 2016 to 2018.
• The decision on whether to discharge or admit a patient can usually be made in less than 24 hours; however, the Medicare benefit does not limit the length of outpatient observation stays. In each of 2014 to 2018, nearly half of outpatient observation visits were longer than 24 hours, including 10 percent that spanned more than 48 hours.
49%
49%
50%
51%
51%
40%
40%
39%
39%
38%
11%
11%
11%
10%
10%
2014
2015
2016
2017
2018
8-23 hours 24-47 hours 48+ hours
42.2
44.4
45.4
45.1
44.9
Outpatient observation visits per 1,000 FFS
beneficairies
A Data Book: Health care spending and the Medicare program, July 2020 73
Chart 6-21. Growth in Medicare FFS hospital outpatient services per capita has slowed, 2010–2018
Note: FFS (fee-for-service). Data for outpatient services include all hospitals paid under the outpatient prospective payment
system, including short-term acute care hospitals, long-term care hospitals, rehabilitation hospitals, and psychiatric hospitals. “Outpatient services per capita” refers to outpatient services per Medicare FFS Part B beneficiary.
Source: MedPAC analysis of hospital outpatient claims data from CMS.
• In 2018, Medicare FFS beneficiaries received approximately 150 million outpatient services at hospitals paid under the outpatient prospective payment system, equivalent to 4.5 outpatient services per Medicare FFS Part B beneficiary (data not shown).
• From 2010 to 2018, the number of Medicare outpatient visits per FFS beneficiary increased 22 percent.
• However, the rate of growth has slowed over time, with outpatient services per FFS beneficiary growing only 0.7 percent in both 2017 and 2018.
3.9%
4.4% 4.3%
3.8%
3.5%
2.0%
1.2%
0.7% 0.7%
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
2010 2011 2012 2013 2014 2015 2016 2017 2018
An
nu
al p
erc
en
tag
e c
han
ge
Calendar year
Cumulative growth(2010–2018):
22%
74 Acute inpatient services
Chart 6-22. Medicare FFS payments to inpatient psychiatric facilities decreased in 2018
Note: FFS (fee-for-service). These fiscal year–incurred data represent only program spending; they do not include beneficiary cost sharing. Spending for inpatient psychiatric care furnished in scatter beds in acute care hospitals (and paid for under the acute care inpatient prospective payment system) is not included in this chart.
Source: CMS Office of the Actuary.
• Medicare pays for inpatient psychiatric facility (IPF) care under the IPF prospective payment system.
• Medicare program spending for FFS beneficiaries’ care in IPFs decreased less than 1 percent per year, on average, from 2010 to 2018.
• However, from 2017 to 2018, Medicare program spending for IPF stays decreased 4.4 percent, reflective of a 6.3 percent decrease in IPF stays (data not shown).
3.83.9 4.0 3.9 4.0 4.0 3.9 3.9
3.7
0
1
2
3
4
5
2010 2011 2012 2013 2014 2015 2016 2017 2018
Do
llars
(in
bil
lio
ns
)
Fiscal year
A Data Book: Health care spending and the Medicare program, July 2020 75
Chart 6-23. A growing share of Medicare-certified inpatient psychiatric facilities are for profit, 2011–2018
Average annual change 2011– 2014– 2017– Type of IPF 2011 2014 2017 2018 2014 2017 2018 All 1,567 1,591 1,601 1,576 0.5% 0.2% –1.6%
Urban 1,234 1,254 1,268 1,246 0.5 0.4 –1.7 Rural 332 336 331 326 0.4 –0.5 –1.5 Freestanding 440 476 516 520 2.7 2.7 0.8 Hospital-based units 1,127 1,115 1,085 1,056 –0.4 –0.9 –2.7 Nonprofit 766 740 736 718 –1.1 –0.2 –2.4 For profit 421 493 522 521 5.4 1.9 –0.2 Government 380 358 343 337 –2.0 –1.4 –1.7
Note: IPF (inpatient psychiatric facility). Data are from facilities that submitted valid Medicare cost reports in the given fiscal year. Components may not sum to totals due to missing data.
Source: MedPAC analysis of hospital cost report data from CMS.
• Between 2011 and 2014, the number of IPFs that filed Medicare cost reports grew, on average, 0.5 percent per year. Similarly, between 2014 and 2017, the supply of IPFs increased slightly, growing, on average, 0.2 percent per year. However, in 2018, the number of IPFs fell by 1.6 percent.
• A growing share of Medicare IPF users receive care in for-profit facilities. Between 2011 and 2014, the number of for-profit IPFs grew 5.4 percent per year, on average. Over the same period, the number of nonprofit IPFs fell more than 1 percent per year, on average. The number of for-profit IPFs continued to grow through 2017, while the number of nonprofit IPFs slightly declined. From 2017 to 2018, the number of for-profit IPFs remained relatively stable, while the number of nonprofit facilities decreased by 2.4 percent.
76 Acute inpatient services
Chart 6-24. Almost three-quarters of Medicare FFS beneficiaries’ stays at IPFs were for psychosis, 2018
MS–DRG Diagnosis Share 885 Psychosis 72.3% 884 Organic disturbances and mental retardation 7.0 057 Degenerative nervous system disorders without MCC 5.8 897 Alcohol/drug abuse or dependency, no rehabilitation, without MCC 4.5 881 Depressive neurosis 3.5 895 Alcohol/drug abuse or dependency with rehabilitation, without MCC 1.7 882 Neurosis except depressive 1.3 880 Acute adjustment reaction and psychosocial dysfunction 0.9 883 Disorders of personality and impulse control 0.7
056 Degenerative nervous system disorders with MCC 0.6 894 Alcohol/drug use—left AMA 0.4 886 Behavioral and developmental disorders 0.2 896 Alcohol/drug abuse or dependency without rehabilitation, with MCC 0.2 876 OR procedure with principal diagnosis of mental illness 0.1 887 Other mental disorders 0.1 081 Nontraumatic stupor and coma without MCC <0.1 080 Nontraumatic stupor and coma with MCC <0.1 Nonpsychiatric MS–DRGs 0.8 Total 100.0
Note: FFS (fee-for-service), IPF (inpatient psychiatric facility), MS–DRG (Medicare severity–diagnosis related group), MCC (major comorbidity or complication), AMA (against medical advice), OR (operating room). Total may not sum to 100 percent due to rounding.
Source: MedPAC analysis of MedPAR data.
• Medicare patients in IPFs are generally assigned 1 of 17 psychiatric MS–DRGs.
• The most frequently occurring IPF diagnosis—accounting for about 72 percent of IPF discharges in 2018—was psychosis. This broad category includes patients with principal diagnoses of schizophrenia, bipolar disorder, and major depression.
• In 2018, the next most common discharge diagnosis, accounting for 7 percent of IPF cases, was organic disturbances and mental retardation.
A Data Book: Health care spending and the Medicare program, July 2020 77
Chart 6-25. The majority of Medicare FFS beneficiaries who received IPF services were under the age of 65, 2018
Share of Share of IPF all IPF users with more Characteristic users than one IPF stay Current eligibility status
Aged 42.6% 30.0% Disabled 57.4 69.9 ESRD only 0.1 0.1 Age
<45 23.1 31.0 45–64 33.6 38.3 65–79 28.8 22.9 80+ 14.4 7.9 All 100.0 27.6 Note: FFS (fee-for-service). IPF (inpatient psychiatric facility), ESRD (end-stage renal disease). The “aged” category includes
beneficiaries ages 65 and older without ESRD. The “disabled” category includes beneficiaries under age 65 without
ESRD. The “ESRD only” category includes beneficiaries with ESRD, regardless of age. Components may not sum to totals due to rounding.
Source: MedPAC analysis of MedPAR data.
• Of Medicare beneficiaries who had at least one IPF stay in 2018, 57.4 percent qualified for Medicare because of a disability. These beneficiaries tend to be younger and poorer than the typical fee-for-service beneficiary.
• Approximately 28 percent of Medicare beneficiaries who used an IPF in 2018 had more than one IPF stay during the year. These beneficiaries were much more likely than all IPF users to be disabled, often because of a psychiatric diagnosis.
7
A Data Book: Health care spending and the Medicare program, July 2020 81
Chart 7-1. Medicare spending per fee-for-service beneficiary on services in the fee schedule for physicians and other health professionals, 2009–2019
Note: Dollar amounts are Medicare spending only and do not include beneficiary cost sharing. The category “disabled” excludes
beneficiaries who qualify for Medicare because they have end-stage renal disease. All beneficiaries age 65 and over are included in the “aged” category.
Source: The annual report of the Boards of Trustees of the Medicare trust funds 2020.
• The fee schedule for physicians and other health professionals includes a broad range of services such as office visits, surgical procedures, and diagnostic and therapeutic services. “Other health professionals” refers to nurse practitioners, physician assistants, physical therapists, and other clinicians. Total fee schedule spending (excluding beneficiary cost sharing) was $73.5 billion in 2019 (data not shown).
• Spending per fee-for-service beneficiary for fee schedule services increased between 2009 and 2011, remained stable between 2011 and 2017, and began growing again after 2017. From 2009 to 2019, spending per beneficiary (across aged and disabled beneficiaries) grew at a cumulative rate of 15 percent.
• Per capita spending for disabled beneficiaries (under age 65) is lower than per capita spending for aged beneficiaries (ages 65 and over). In 2019, for example, per capita spending for disabled beneficiaries was $1,893 compared with $2,227 for aged beneficiaries. However, spending per capita grew much faster for disabled beneficiaries than aged beneficiaries between 2009 and 2019 (20 percent vs. 13 percent, respectively).
1,963
2,142 2,125 2,124 2,1042,227
1,581
1,732 1,764 1,8041,749
1,893
0
200
400
600
800
1,000
1,200
1,400
1,600
1,800
2,000
2,200
2,400
2,600
2009 2011 2013 2015 2017 2019
Sp
en
din
g p
er
ben
efi
cia
ry (
do
llars
)
Aged
Disabled
82 Ambulatory care
Chart 7-2. Physician fee schedule–allowed charges by type of service, 2018
Total allowed charges in 2018 = $94.7 billion
Source: MedPAC analysis of the Carrier Standard Analytic File for 100 percent of beneficiaries.
• In 2018, allowed charges for physician fee schedule services totaled $94.7 billion. Allowed charges include both program spending and beneficiary cost sharing.
• In 2018, about half of all allowed charges were for evaluation and management (E&M) services.
• Within the E&M category, about half of allowed charges were for office/outpatient visits. The remaining allowed charges within the E&M category were for various types of services that occurred across a broad range of settings, including hospital inpatient departments, emergency departments, and nursing facilities (data not shown).
Evaluation and management
50.4%
Other procedures22.6%
Imaging11.0%
Major procedures7.5%
Tests5.1%
Anesthesia2.9%
Other0.5%
A Data Book: Health care spending and the Medicare program, July 2020 83
Chart 7-3. Total encounters per beneficiary increased and mix of clinicians furnishing them changed from 2013 to 2018
Specialty category
Encounters per beneficiary Percent change in
encounters per beneficiary
2013 2018 Average annual Total
Total (all clinicians) 20.8 21.9 1.0% 5.0%
Primary care physicians 4.1 3.6 –2.9 –13.7
Specialists 12.5 12.8 0.4 2.0
APRNs/PAs 1.3 2.2 11.5 72.1
Other practitioners 2.8 3.3 2.8 15.1
Note: APRN (advanced practice registered nurse), PA (physician assistant). We define “encounters” as unique combinations of
beneficiary identification numbers, claim identification numbers (for paid claims), and national provider identifiers of the clinicians who billed for the service. Figures may not sum to totals due to rounding. Figures do not account for “incident to” billing, meaning, for example, that encounters with APRNs/PAs that are billed under Medicare’s “incident to” rules are
included in the physician totals. We use the number of fee-for-service beneficiaries enrolled in Part B to define encounters per beneficiary.
Source: MedPAC analysis of the Carrier Standard Analytic File for 100 percent of beneficiaries and 2019 annual report of the
Boards of Trustees of the Medicare trust funds.
• Encounters measure beneficiary interactions with clinicians. For example, if a physician billed for an office visit and an X-ray on the same claim, we count that as one encounter.
• The number of encounters per beneficiary grew 1 percent per year from 2013 to 2018, suggesting stable access to care.
• Encounters with specialist physicians accounted for a majority of all encounters and grew modestly from 2013 to 2018.
• In contrast, encounters with APRNs or PAs grew rapidly from 2013 to 2018, and encounters with primary care physicians declined substantially. These changes continue a longer term trend of declines in services billed by primary care physicians and rapid increases in services billed by APRNs and PAs (data not shown).
• The decline in encounters with primary care physicians occurred across a broad range of services, including evaluation and management services, tests, procedures, and imaging services (data not shown).
84 Ambulatory care
Chart 7-4. Medicare beneficiaries’ ability to get timely appointments with physicians was comparable with privately insured individuals, 2016–2019
Medicare (ages 65 and older) Private insurance (ages 50–64)
Survey question 2016 2017 2018 2019 2016 2017 2018 2019
Unwanted delay in getting an appointment: Among those who needed an appointment, “How often did you have to wait longer than you wanted to get a doctor’s appointment?”
For routine care
Never 68%b 73%a 70%ab 72% 67%b 69%ab 64%ab 74%
Sometimes 22 20a 20a 20 23b 22 ab 26ab 19
Usually 4b 3 5b 3 5 4 5 4
Always 3 3 3a 3 4b 3 4ab 3
For illness or injury
Never 79a 80a 79a 80 75ab 76ab 74ab 81
Sometimes 16a 15a 15a 14 19ab 18ab 19ab 15
Usually 2a 2 2 2 3ab 2 3b 2
Always 2a 1a 2 2 3ab 2a 2 1
Note: Numbers may not sum to 100 percent due to rounding and to missing responses (“Don’t Know” or “Refused”) not being
presented. Overall sample sizes for each group (Medicare and privately insured) were approximately 4,000 in all years.
Sample sizes for individual questions varied.
aStatistically significant difference (at a 95 percent confidence level) between the Medicare and privately insured samples
in the given year. bStatistically significant difference (at a 95 percent confidence level) from 2019 within the same insurance coverage
category.
Source: MedPAC-sponsored annual telephone surveys conducted 2016–2019.
• Most Medicare beneficiaries have one or more doctor appointments in a given year. Their ability to schedule timely appointments is one indicator of access that we examine.
• Medicare beneficiaries (ages 65 and older) report similar access to physicians for appointments as compared with privately insured individuals ages 50 to 64. For example, in 2019, 72 percent of Medicare beneficiaries reported that they never had to wait longer than they wanted for routine care, and 80 percent reported the same for illness or injury care. Medicare beneficiaries’ ability to obtain either type of care when needed was statistically no different compared with privately insured individuals (the comparable rates for privately insured individuals were 74 percent for routine care and 81 percent for illness or injury care).
• Appointment scheduling for illness and injury is better than for routine care appointments for both Medicare beneficiaries and privately insured individuals.
A Data Book: Health care spending and the Medicare program, July 2020 85
Chart 7-5. Medicare and privately insured patients who were looking for a new physician reported more difficulty finding one in primary care, 2016–2019
Medicare (ages 65 and older) Private insurance (ages 50–64)
Survey question 2016 2017 2018 2019 2016 2017 2018 2019
Looking for a new physician “In the past 12 months, have you tried to get a new …?” (Percent
answering “Yes”)
Primary care physician 8%a 9%a 10%b 8% 10%a 11%ab 10% 9%
Specialist 18 17a 19ab 17 18b 20ab 21ab 15
Getting a new physician: Among those who tried to get an appointment with a new physician, “How much of a problem was it finding a primary care doctor/specialist who would treat you? Was it … ?”
Primary care physician
No problem 64 69a 71 72a 63 59a 67 62a
Small problem 15 13 13 13a 16 18 16 20a
Big problem 20 14a 14 14 20 22a 16 17
Specialist
No problem 82 83 84 85a 79 81 80 79a
Small problem 10b 11b 7 6a 9 11 9 11a
Big problem 8a 5ab 8 8 11a 8a 10 9
Note: Numbers may not sum to 100 percent due to rounding and to missing responses (“Don’t Know” or “Refused”) not being
presented. Overall sample sizes for each group (Medicare and privately insured) were approximately 4,000 in all years. Sample sizes for individual questions varied.
aStatistically significant difference (at a 95 percent confidence level) between the Medicare and privately insured samples in the
given year. bStatistically significant difference (at a 95 percent confidence level) from 2019 within the same insurance coverage category.
Source: MedPAC-sponsored annual telephone surveys, conducted 2016–2019.
• In 2019, only 8 percent of Medicare beneficiaries and 9 percent of privately insured individuals reported
looking for a new primary care physician. This finding suggests that most people were either satisfied with their current physician or did not need to look for one.
• In 2019, Medicare beneficiaries and privately insured individuals were more likely to report problems finding a new primary care physician than a new specialist.
• Of the 8 percent of Medicare beneficiaries who looked for a new primary care physician in 2019, 27 percent reported problems finding one: 14 percent reported their problem as “big,” and 13 percent reported their problem as “small.” Although this finding means that less than 3 percent of the total Medicare population reported problems finding a primary care physician, the Commission is concerned about the continuing pattern of greater problems accessing primary care than specialty care.
• Of the 9 percent of privately insured individuals who looked for a new primary care physician in 2019, 37 percent reported problems finding one: 17 percent reported their problem as “big,” and 20 percent reported their problem as “small.”
86 Ambulatory care
Chart 7-6. Medicare beneficiaries’ access to physician care was comparable with privately insured individuals, and minorities in both groups reported unwanted delays more frequently, 2019
Medicare (ages 65 and older) Private insurance (ages 50–64)
Survey question All White Minority All White Minority
Unwanted delay in getting an appointment: Among those who needed an appointment, “How often did you have to wait longer than you wanted to get a doctor’s appointment?”
For routine care
Never 72% 74%b 68%b 74% 76%b 68%b
Sometimes 20 19 22 19 18b 22b
Usually 3 3 3a 4 3b 6ab
Always 3 2 3 3 2 3
For illness or injury
Never 80 82b 76b 81 83b 77b
Sometimes 14 13b 18b 15 14b 18b
Usually 2 2 3 2 2 2
Always 2 2 1 1 1b 3b
Note: Numbers may not sum to 100 percent due to rounding and to missing responses (“Don’t Know” or “Refused”) not being
presented. Overall sample size for each group (Medicare and privately insured) was approximately 4,000 in 2019. Sample size for individual questions varied.
aStatistically significant difference (at a 95 percent confidence level) between the Medicare and privately insured
populations in the given category.
bStatistically significant difference (at a 95 percent confidence level) by race within the same insurance category.
Source: MedPAC-sponsored telephone surveys conducted in 2019.
• In 2019, Medicare beneficiaries (ages 65 and older) reported similar access to physicians for appointments in comparison with privately insured individuals ages 50 to 64.
• Access varied by race, with minorities more likely than Whites to report access problems in both insurance categories. For example, in 2019, 82 percent of White Medicare beneficiaries reported “never” having to wait longer than they wanted to get an appointment for an illness or injury compared with 76 percent of minority beneficiaries.
A Data Book: Health care spending and the Medicare program, July 2020 87
Chart 7-7. Minorities in Medicare and with private insurance were more likely to report problems finding a new specialist, 2019
Medicare (ages 65 and older) Private insurance (ages 50–64)
Survey question All White Minority All White Minority
Looking for a new physician: “In the past 12 months, have you tried to get a new …?”
Primary care physician 8% 8% 8% 9% 9% 9%
Specialist 17 18b 14b 15 16 13
Getting a new physician: Among those who tried to get an appointment with a new physician, “How much of a problem was it finding a primary care doctor/specialist who would treat you? Was it … ?”
Primary care physician
No problem 72a 74 66 62a 65 56
Small problem 13a 12 14 20a 19 23
Big problem 14 12 20 17 16 20
Specialist
No problem 85a 88ab 75b 79a 81ab 72b
Small problem 6a 6 9 11a 9b 18b
Big problem 8 7b 16b 9 9 10
Note: Numbers may not sum to 100 percent due to rounding and to missing responses (“Don’t Know” or “Refused”) not being
presented. Overall sample size for each group (Medicare and privately insured) was approximately 4,000 in 2019. Sample
size for individual questions varied.
aStatistically significant difference (at a 95 percent confidence level) between the Medicare and privately insured
populations in the given category.
bStatistically significant difference (at a 95 percent confidence level) by race within the same insurance category.
Source: MedPAC-sponsored telephone surveys conducted in 2019.
• Overall, Medicare beneficiaries reported fewer problems finding a new primary care physician or specialist than those with private insurance.
• Among those looking for a specialist, minorities were more likely than Whites to report problems finding one. This pattern held for Medicare beneficiaries and for privately insured individuals ages 50 to 64.
88 Ambulatory care
Chart 7-8. Changes in physicians’ professional liability insurance premiums, 2012–2019
Note: Bars represent a four-quarter moving average percentage change.
Source: CMS, Office of the Actuary. Data are from CMS’s Professional Liability Physician Premium Survey.
• Medicare’s fee schedule for physicians and other health professionals includes payments to clinicians that are intended to cover the relative cost of professional liability insurance (PLI) premiums. Payments for PLI account for 4.3 percent of total payments under the fee schedule (data not shown).
• Changes in the PLI premiums paid by physicians and other health professionals reflect a cyclical pattern, alternating between periods of low premiums (characterized by high investment returns for insurers and vigorous competition) and high premiums (characterized by declining investment returns and market exit).
• Premiums grew slowly from the second quarter of 2012 through the first quarter of 2014, declined from the second quarter of 2014 through the third quarter of 2018, and began increasing again in the first quarter of 2019.
-0.2
1.2
1.6
1.0
0.3
-0.5
-1.0
-1.2-1.1
-0.7
-0.5-0.6
-0.5
-0.1
0.2
0.6
-2
-1
0
1
2
2012 2013 2014 2015 2016 2017 2018 2019
Perc
en
t ch
an
ge
A Data Book: Health care spending and the Medicare program, July 2020 89
Chart 7-9. Spending on hospital outpatient services covered under the outpatient PPS, 2009–2019
Note: PPS (prospective payment system). Spending amounts are for services covered by the Medicare outpatient PPS. They do
not include services paid on separate fee schedules (e.g., ambulance services and durable medical equipment) or those paid on a cost basis (e.g., corneal tissue acquisition and flu vaccines) or payments for clinical laboratory services, except
those packaged into payment bundles. *Estimated figures.
Source: CMS, Office of the Actuary.
• The Office of the Actuary estimates that spending under the outpatient PPS was $72.7 billion in 2019 ($58.7 billion in program spending, $14.0 billion in beneficiary copayments). We estimate that the outpatient PPS accounted for about 7 percent of total Medicare program spending in 2019 (data not shown).
• Overall spending by Medicare and beneficiaries on hospital outpatient services covered under the outpatient PPS from calendar years 2009 to 2019 increased by 114 percent, an average of 7.9 percent per year. The Office of the Actuary projects continued growth in total spending, averaging 10.4 percent per year from 2019 to 2021 (data not shown).
• Beneficiary cost sharing under the outpatient PPS includes the Part B deductible and coinsurance for each service. Under the outpatient PPS, beneficiary cost sharing was about 19 percent in 2019 (data not shown).
26.3 28.6 31.1 33.9 36.341.2
44.8 47.9 51.4 54.958.77.7
8.28.7
9.210.2
11.311.5
12.012.5
13.714.0
0
10
20
30
40
50
60
70
80
2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019*
Do
llars
(in
bil
lio
ns)
Beneficiary cost sharing
Program payments
90 Ambulatory care
Chart 7-10. Most hospitals provide outpatient services
Share offering
Acute care Outpatient Outpatient Emergency Year hospitals services surgery services
2008 3,607 94% 87% N/A 2010 3,518 95 90 N/A 2012 3,483 95 91 93% 2014 3,429 96 92 93 2016 3,370 96 93 93 2018 3,301 96 93 90 2019 3,245 96 93 91
Note: N/A (not applicable). We list emergency services from 2008 through 2010 as “N/A” because the data source we used in
this chart changed the variable for identifying hospitals’ provision of emergency services. We believe this change in
variable definition makes it appear that the share of hospitals providing emergency services increased sharply from 2010 to 2012, but we question whether such a large increase actually occurred. This chart includes services provided or arranged by acute care short-term hospitals and excludes long-term, Christian Science, psychiatric, rehabilitation,
children’s, critical access, and alcohol/drug hospitals.
Source: Medicare Provider of Services files from CMS.
• The number of hospitals that furnish services under Medicare’s outpatient prospective payment system has declined slowly since 2008, from 3,607 in 2008 to 3,245 in 2019.
• The share of hospitals providing outpatient services remained stable, and the share offering outpatient surgery steadily increased from 2008 through 2014 and has remained stable since then. The share offering emergency services declined slightly from 2016 to 2018.
A Data Book: Health care spending and the Medicare program, July 2020 91
Chart 7-11. Payments and volume of services under the Medicare hospital outpatient PPS, by type of service, 2018
Payments Volume
Note: PPS (prospective payment system), E&M (evaluation and management). “Payments” include both program spending and beneficiary cost sharing. We grouped services into the following categories, according to the Berenson-Eggers Type of Service codes developed by CMS: evaluation and management, procedures, imaging, and tests. “Pass-through drugs”
and “separately paid drugs/blood products” are classified by their payment status indicator. The percentages in the “volume” figure do not sum to 100 due to rounding.
Source: MedPAC analysis of standard analytic file of outpatient claims for 2018.
• Hospitals provide many types of services in their outpatient departments, including emergency and clinic visits, imaging and other diagnostic services, laboratory tests, and ambulatory surgery.
• The payments for services are distributed differently from volume. For example, in 2018, procedures accounted for 47 percent of payments but only 15 percent of volume.
• Procedures (e.g., endoscopies, surgeries, and skin and musculoskeletal procedures) accounted for the greatest share of payments for services (47 percent) in 2018, followed by evaluation and management services (19 percent), separately paid drugs and blood products (18 percent), and imaging services (12 percent).
• Payments for separately payable drugs and blood products and pass-through drugs have increased in relation to other categories in the outpatient PPS, increasing from 15 percent of total outpatient PPS spending in 2013 to 21 percent of total outpatient PPS spending in 2018 (data not shown). Pass-through drugs are new drugs that have been approved by the FDA; were not paid under Medicare’s hospital outpatient payment system before January 1, 1997; and have been determined to have costs that are not insignificant in relation to the outpatient PPS payment rate for the applicable service. Statute allows drugs to have pass-through status for two to three years.
Tests1%
Pass-through drugs3%
Procedures47%
Imaging12%
E&M19%
Separately paid drugs/
blood products
18%
Tests2%
Pass-through drugs5%
Procedures15%
Imaging9%
E&M 16%
Separately paid
drugs/blood products
54%
92 Ambulatory care
Chart 7-12. Hospital outpatient services with the highest Medicare expenditures, 2018
Share of Medicare Volume Payment APC title expenditures (thousands) rate Total 52%
All emergency visits 7 13,010 $337
Clinic visits 6 32,462 114
Comprehensive observation services 5 1,430 2,350
Level 3 endovascular procedures 3 189 10,510
Level 5 musculoskeletal procedures 2 135 10,123
Level 3 electrophysiologic procedures 2 68 18,516
Level 3 drug administration 2 6,544 191
Level 2 ICD and similar procedures 2 41 30,962
Level 4 musculoskeletal procedures 2 224 5,606
Level 1 endovascular procedures 2 355 2,814
Level 3 radiation therapy 1 1,843 522
Level 2 imaging without contrast 1 8,413 114
Level 4 imaging without contrast 1 1,828 487
Level 1 intraocular procedures 1 458 1,921
Level 1 laparoscopy and related procedures 1 193 4,489
Level 3 nuclear medicine and related services 1 706 1,203
Level 2 lower GI procedures 1 996 936
Level 3 imaging without contrast 1 3,498 232
Level 5 urology and related services 1 204 3,706
Level 4 drug administration 1 2,373 298
Level 1 imaging with contrast 1 2,802 253
Level 3 pacemaker and similar procedures 1 69 9,748
Level 1 upper GI procedures 1 963 743
Level 4 endovascular procedures 1 40 16,020
Level 4 neurostimulator and related procedures 1 23 27,892
Level 2 excision/biopsy/incision and drainage 1 451 1,348
Level 3 vascular procedures 1 249 2,493
Average APC 585 174 Note: APC (ambulatory payment classification), ICD (implantable cardioverter-defibrillator), GI (gastrointestinal). The payment
rate for “all emergency visits” is a weighted average of payment rates for 10 emergency visit APCs (not listed on this chart). The shares of payments for the 27 APC categories do not add to the total share of payments (52 percent) because of rounding. The average APC figures in the last line represent averages for all APCs.
Source: MedPAC analysis of 100 percent analytic files of outpatient claims for calendar year 2018.
• Although the outpatient prospective payment system covers thousands of services,
expenditures are concentrated in a few categories that have high volume, high payment rates, or both.
A Data Book: Health care spending and the Medicare program, July 2020 93
Chart 7-13. Separately payable drugs have increased as a share of total spending in the outpatient prospective payment system, 2013–2018
Note: OPPS (outpatient prospective payment system). The dotted line segment indicates our estimate of the share of total
OPPS spending in 2018 that would have been attributable to separately payable drugs if CMS had not implemented a policy in 2018 that substantially reduced the OPPS payment rates for drugs obtained through the 340B Drug Pricing
Program. Source: MedPAC analysis of hospital outpatient standard analytic claims files from 2013 through 2018.
• The OPPS packages the cost of most drugs into the payment for the related services. However, the OPPS has two programs that provide separate payment for higher cost drugs: the pass-through program, which is focused on drugs that are new to the market, and the program for separately payable non-pass-through (SPNPT) drugs, which is focused on drugs that have been established in the drug market. Pass-through drugs can hold that status for two to three years, after which they can become SPNPT drugs. Most SPNPT drugs were previously pass-through drugs.
• Separately payable drugs have become an increasingly larger share of OPPS spending, increasing from 15.5 percent in 2013 to 20.5 percent in 2018.
• The share of OPPS spending attributable to separately payable drugs decreased slightly from 2013 to 2014 and increased only slightly from 2017 to 2018. The decrease from 2013 to 2014 in separately payable drugs’ share of spending was the result of an unusually large increase in total OPPS spending caused by a change in policy that allowed for many clinical lab tests to be paid under the OPPS rather than the clinical lab fee schedule. The small increase from 2017 to 2018 was the result of a policy implemented by CMS that substantially decreased the payment rates for SPNPT drugs that hospitals obtained through the 340B Drug Pricing Program. Without that policy, we estimate that separately payable drugs would have been 22.7 percent of OPPS spending in 2018.
15.514.7
16.3
18.4
20.4
20.5
22.7
0
5
10
15
20
25
30
2013 2014 2015 2016 2017 2018
Sep
ara
tely
payab
le d
rug
s a
s a
sh
are
of
OP
PS
sp
en
din
g
Actual
Estimated without CMSpolicy change
94 Ambulatory care
Chart 7-14. Number of Medicare-certified ASCs increased by 10 percent, 2012–2018 2012 2013 2014 2015 2016 2017 2018
Medicare payments (billions of dollars) $3.6 $3.7 $3.8 $4.1 $4.3 $4.6 $4.9 New centers (during year) 179 179 190 170 168 215 224 Closed or merged centers (during year) 114 120 123 108 100 94 78 Net total number of centers (end of year) 5,194 5,253 5,320 5,382 5,450 5,571 5,717 Net percent growth in number of centers 1.3% 1.1% 1.3% 1.2% 1.3% 2.2% 2.6% Share of all centers that are: For profit 94 95 95 95 95 95 95 Nonprofit 4 4 4 4 4 4 4 Government 2 2 2 2 2 2 2 Urban 93 93 93 93 93 93 93 Rural 7 7 7 7 7 7 7
Note: ASC (ambulatory surgical center). Medicare payments include program spending and beneficiary cost sharing for ASC
facility services. Some figures do not match to Chart 7-15 in our 2019 data book because CMS updated the Provider of
Services file. Some totals may not sum to 100 percent due to rounding. Source: MedPAC analysis of Provider of Services file from CMS 2018. Payment data are from CMS, Office of the Actuary.
• ASCs are distinct entities that furnish ambulatory surgical services not requiring an overnight stay in a hospital. The most common ASC procedures are cataract removal with lens insertion, upper gastrointestinal endoscopy, colonoscopy, and nerve procedures.
• Total Medicare payments per fee-for-service (FFS) Medicare beneficiary for ASC services increased by approximately 5 percent per year, on average, from 2012 through 2018 (data not shown). Payments per FFS beneficiary served in an ASC grew by 4.9 percent per year during this period. From 2017 to 2018, total payments rose by 6.4 percent, and payments per beneficiary grew by 7.4 percent (per beneficiary data not shown).
• The number of Medicare-certified ASCs grew at an average annual rate of 1.6 percent from 2012 through 2018. In this same period, an annual average of 189 new facilities entered the market, while an average of 105 closed or merged with other facilities.
• Compared with earlier years (not shown), the number of ASCs grew slowly from 2012 through 2018. The slower growth may reflect the substantially higher rates that Medicare pays for ambulatory surgical services provided in hospital outpatient departments than in ASCs and the significant increase in hospital employment of physicians.
8
A Data Book: Health care spending and the Medicare program, July 2020 97
Chart 8-1. The number of post-acute care providers decreased slightly in 2019
Average
annual
percent Percent
change change
2015 2016 2017 2018 2019 2015−2019 2018−2019
Home health agencies 12,346 12,204 11,844 11,783 11,356 –2.1% –3.6% Inpatient rehabilitation facilities 1,182 1,188 1,178 1,170 1,152 –0.6 –1.5 Long-term care hospitals 426 423 411 386 371 –3.4 –3.9 Skilled nursing facilities 15,223 15,263 15,277 15,230 15,114 –0.2 –0.8
Note: The skilled nursing facility count does not include swing beds.
Source: MedPAC analysis of data from the Provider of Services files from CMS.
• The number of home health agencies has been declining since 2015 after several years of substantial growth (data not shown). The decline in agencies was concentrated in Texas and Florida, two states that saw considerable growth after the implementation of the home health prospective payment system in October 2000.
• The supply of inpatient rehabilitation facilities (IRFs) has been declining slightly since 2015, though the rate of change picked up between 2018 and 2019. Most IRFs are distinct units in acute care hospitals; about one-quarter are freestanding facilities. However, because freestanding IRFs tend to have more beds, they account for about half of Medicare discharges from IRFs.
• After peaking in 2012 (data not shown), the number of long-term care hospitals (LTCHs) has
decreased. The number of LTCHs declined more rapidly after the implementation of a new
“dual payment-rate structure” that reduces payments for certain Medicare discharges from
LTCHs beginning in fiscal year 2016.
• The total number of skilled nursing facilities rose between 2015 and 2017, then decreased less than 1 percent per year between 2017 and 2019.
98 Post-acute care
Chart 8-2. Medicare fee-for-service spending for post-acute care expenditures was relatively stable from 2010 to 2018
Note: These calendar year‒incurred data represent only program spending; they do not include beneficiary cost sharing.
Source: CMS Office of the Actuary 2020.
• Aggregate fee-for-service (FFS) spending on post-acute care (PAC) has remained stable since 2012, in part because of expanded enrollment in managed care under Medicare Advantage (Medicare Advantage spending is not included in this chart). However, spending growth has varied by PAC sector.
• FFS spending on skilled nursing facilities increased sharply in 2011, reflecting CMS’s adjustment for the implementation of the new case-mix groups (resource utilization groups, version IV). Once CMS established that the adjustment it made was too large, it lowered the adjustment, and spending dropped in 2012. Overall, spending on SNF care and home health care was relatively stable between 2012 and 2018, decreasing slightly in the latter part of the period.
• FFS spending on inpatient rehabilitation facilities (IRFs) has increased steadily over the past decade. In all, spending on IRFs increased 37 percent between 2010 and 2018.
• FFS spending on long-term care hospitals (LTCHs) has decreased by 24 percent since 2015, largely due to the implementation of the dual payment-rate structure that reduced payments for certain LTCH cases.
58.661.3
58.3 58.8 59.2 60.4 59.7 58.8 58.5
27.831.1
28.2 28.6 29.0 29.6 28.9 28.5 28.1
19.6 18.6 18.2 18.1 18.0 18.4 18.4 18.1 18.1
6.1 6.4 6.7 6.9 7.2 7.5 7.7 8.1 8.4
5.0 5.2 5.3 5.2 5.0 5.0 4.7 4.1 3.80
10
20
30
40
50
60
70
2010 2011 2012 2013 2014 2015 2016 2017 2018
Do
llars
(in
bil
lio
ns)
Calendar year
All post-acute care
Skilled nursing facilities
Home health agencies
Inpatient rehabilitation facilities
Long-term care hospitals
A Data Book: Health care spending and the Medicare program, July 2020 99
Chart 8-3. Freestanding SNFs and for-profit SNFs accounted for the majority of facilities, Medicare stays, and Medicare spending in 2012 and 2018
Medicare payments Facilities Medicare-covered stays (billions)
Type of SNF 2012 2018 2012 2018 2012 2018 Totals 14,938 15,042 2,396,548 2,191,246 $26.2 $25.4 Freestanding 95% 96% 94% 96% 97% 97% Hospital based 5 4 6 4 3 3 Urban 70 73 82 84 84 85 Rural 30 27 18 16 16 15 For profit 70 71 71 71 75 74 Nonprofit 25 23 25 25 21 22 Government 5 6 3 4 3 4
Note: SNF (skilled nursing facility). Totals may not sum to 100 percent due to rounding and missing values. The spending
amount included here is lower than that reported by the Office of the Actuary, and the count of SNFs is slightly lower than
what is reported in CMS’s Survey and Certification Providing Data Quickly system. Source: MedPAC analysis of the Provider of Services and Medicare Provider Analysis and Review files, 2012 and 2018.
• In 2018, freestanding facilities accounted for 96 percent of stays and 97 percent of Medicare’s payments.
• Urban facilities accounted for 73 percent of facilities, 84 percent of stays, and 85 percent of Medicare payments in 2018.
• In 2018, for-profit facilities accounted for 71 percent of facilities and stays and 74 percent of Medicare payments.
100 Post-acute care
Chart 8-4. SNF admissions and stays continued to decline in 2018
Percent change Volume measure 2014 2016 2017 2018 2017‒2018
Covered admissions per 1,000 FFS beneficiaries 68.3 65.9 64.6 62.5 –3.3% Covered days per 1,000 1,843 1,693 1,623 1,559 –3.9 FFS beneficiaries Covered days per admission 27.0 25.7 25.1 25.0 –0.4 Note: SNF (skilled nursing facility), FFS (fee-for-service). Data include 50 states and the District of Columbia. Source: Calendar year data from CMS, Office of Information Products and Data Analytics 2020.
• In 2018, 4 percent of beneficiaries enrolled in Medicare fee-for-service used SNF services, down slightly from 2011 (data not shown).
• Between 2017 and 2018, SNF admissions per 1,000 FFS beneficiaries decreased 3.3 percent. The decline is consistent with a decline in FFS per capita inpatient hospital stays that were three days or longer and therefore qualified for Medicare coverage of SNF care (data not shown).
• During the same period, covered days per admission declined 0.4 percent to 25 days, so there were fewer covered days per 1,000 beneficiaries.
A Data Book: Health care spending and the Medicare program, July 2020 101
Chart 8-5. Freestanding SNF Medicare margins remained high in 2018
2012 2014 2015 2016 2017 2018 All 14.1% 12.8% 12.6% 11.6% 11.3% 10.3% Rural 13.3 10.9 10.9 9.9 9.7 8.2 Urban 14.2 13.1 13.0 11.9 11.5 10.7 Nonprofit 5.7 4.2 4.4 2.3 1.7 0.5 For profit 16.3 15.2 15.1 14.2 13.7 13.0 Note: SNF (skilled nursing facility).
Source: MedPAC analysis of freestanding SNF cost reports 2012–2018.
• Though lower than in recent years, the aggregate Medicare margin for freestanding SNFs in 2018 exceeded 10 percent for the 19th consecutive year (not all years are shown). After reaching over 21 percent in 2011 (data not shown), the margins have declined primarily because current law requires annual market basket increases to payments to be offset by a productivity adjustment.
• In 2018, on average, urban facilities had higher Medicare margins than rural facilities. For-profit SNFs had considerably higher Medicare margins than nonprofit SNFs, reflecting their larger size and lower cost growth.
• In 2018, the average total margin (the margin across all payers and all lines of business) for freestanding facilities was –0.3, the first year that it was negative since 1999 (data not shown).
102 Post-acute care
Chart 8-6. Cost and payment differences explain variation in Medicare margins for freestanding SNFs in 2018 Highest margin Lowest margin Ratio of highest quartile quartile quartile to Characteristic (n = 3,318) (n = 3,318) lowest quartile Cost measures
Standardized cost per day $278 $410 0.68 Standardized cost per discharge $11,392 $14,506 0.79 Average daily census (patients) 88 65 1.34 Revenue measures
Medicare payment per day $530 $458 1.16 Medicare payment per discharge $22,554 $15,730 1.43 Share of days in intensive therapy 89% 81% 1.10 Share of medically complex days 3 3 1.00 Medicare share of facility revenue 22 12 1.83 Average length of stay (days) 41 34 1.20 Medicaid share of days 66 57 1.16 Patient characteristics
Case-mix index 1.42 1.32 1.08 Share of dual-eligible beneficiaries 51% 36% 1.42 Share of minority beneficiaries 15 5 3.00 Share of very old beneficiaries 26 33 0.79 Facility mix
Share for profit 85% 55% N/A Share urban 81 70 N/A
Note: SNF (skilled nursing facility), N/A (not applicable). Values shown are medians for the quartile. Highest margin quartile SNFs
were in the top 25 percent of the distribution of Medicare margins. Lowest margin quartile SNFs were in the bottom 25
percent of the distribution of Medicare margins. “Standardized cost per day” includes Medicare costs adjusted for differences in area wages and the case mix (using the nursing component’s relative weights) of Medicare beneficiaries. “Days in intensive therapy” are days classified into ultra-high and very high rehabilitation case-mix groups. “Very old beneficiaries” are
85 years or older. “Medically complex days” are those assigned to clinically complex or special-care case-mix groups. Quartile figures presented in the table are rounded, but the ratio column was calculated using unrounded data.
Source: MedPAC analysis of freestanding SNF claims and cost reports 2018.
• Medicare margins varied widely across freestanding SNFs. One-quarter of SNFs had Medicare margins at or below –0.7 percent, and one-quarter of facilities had Medicare margins at or above 19.7 percent (data not shown).
• High-margin SNFs had lower costs per day (32 percent lower costs than low-margin SNFs), after adjusting for wage and case-mix differences, and higher payment per day (16 percent).
• Facilities with the highest Medicare margins had higher case-mix indexes, higher shares of beneficiaries who were dually eligible for Medicare and Medicaid, and higher shares of minority beneficiaries.
A Data Book: Health care spending and the Medicare program, July 2020 103
Chart 8-7. Financial performance of relatively efficient SNFs in 2018 reflects a combination of lower cost per day and higher payment per day
Relatively Other efficient SNFs SNFs
Performance in 2018
Community discharge rate 52% 41% Readmission rate 9% 10%
Standardized cost per day $304 $331 Medicare revenue per day $530 $482
Medicare margin 16.9% 9.9% Total margin 2.0% 0.3%
Facility case-mix index 1.44 1.36 Medicare average length of stay 30 days 37 days Occupancy rate 88% 84% Average daily census 98 78
Share of ultra-high therapy days 69% 56% Share of medically complex days 4% 4%
Medicaid share of facility days 58% 63%
Share urban 85% 68%
Share for profit 79% 67% Note: SNF (skilled nursing facility). The analysis includes 11,551 freestanding facilities. SNFs were defined as “relatively
efficient” by their cost per day measure (2015–2017) and two quality measures (community discharge and readmission rates) for the same period (2015–2017). Relatively efficient SNFs were those in the best third of the distribution of any one measure and not in the bottom third on any measure in each of three years. Eight percent of SNFs qualified as relatively
efficient. Costs per day were standardized for differences in case mix (using the nursing component relative weights) and wages. Rates of risk-adjusted community discharge and readmission for patients with potentially avoidable conditions during the SNF stay are quality measures and were calculated for all facilities with at least 25 stays. “Ultra-high therapy
days” include days with at least 720 minutes per week of therapy. “Medically complex days” are those assigned to clinically complex or special-care case-mix groups.
Source: MedPAC analysis of quality measures and Medicare cost report data for 2015–2018.
• “Relatively efficient SNFs” are defined as those consistently providing relatively low-cost and high-quality care compared with other SNFs. Compared with other SNFs in 2018, relatively efficient SNFs furnished considerably higher quality (higher discharge to community rates and lower readmission rates) and had costs per day that were 8 percent lower.
• Compared with other SNFs in 2018, relatively efficient SNFs treated the same share of medically complex patients, had a higher share of ultra-high therapy days, shorter stays, higher occupancy rates, and higher average daily censuses.
104 Post-acute care
Chart 8-8. SNFs improved on some measures but not others from 2012 to 2018
Measure 2012 2014 2016 2018
Discharged to the community 35.7% 37.7% 39.6%
41.4%
Potentially avoidable readmissions
During SNF stay 11.4 10.8 10.8 10.6
During 30 days after discharge from SNF 5.7 5.7 5.8 5.9
Rate of improvement in one or more mobility ADLs 43.6 43.5 43.6 43.9
Rate of no decline in mobility 87.2 87.1 87.1 87.2
Note: SNF (skilled nursing facility), ADL (activity of daily living). High rates of discharge to the community indicate better qual ity. High readmission rates indicate worse quality. All rates were risk adjusted. The rate of improvement in mobility ADLs is the average of the rates of improvement in bed mobility, transfer, and ambulation, weighted by the number of stays
included in each measure. Stays with improvement in one, two, or three mobility ADLs are counted in the improvement measures. “Rate of no decline in mobility” is the share of stays with no decline in any of the three ADLs. Rates are the average of facility rates and calculated for all facilities with 25 or more stays, except the rate of potentially avoidable
readmissions during the 30 days after discharge, which is reported for all facilities with 20 or more stays. Measures exclude hospital-based swing-bed units.
Source: MedPAC analysis of Medicare claims and Minimum Data Set data for 2012–2018.
• Quality measures for SNFs draw on two sources: claims for payment submitted by SNFs and patient assessment data collected by SNFs. Given evidence that the patient assessment information reported by inpatient rehabilitation facilities and home health agencies may reflect financial considerations, these measures should be interpreted carefully.
• Rates of claims-based, risk-adjusted community discharge (discharged home with or without home health care) and potentially avoidable readmission during the SNF stay improved between 2012 and 2018. A greater share of beneficiaries was discharged to the community (41.4 percent compared with 35.7 percent). A lesser share of beneficiaries was readmitted to an acute care hospital during the SNF stay (10.6 percent compared with 11.4 percent). The share of beneficiaries readmitted to an acute care hospital in the 30 days after discharge from the SNF has increased slightly since 2012, to 5.9 percent in 2018.
• Both readmission rates include only patients readmitted to a hospital with the principal diagnosis of a potentially avoidable condition. The 13 potentially avoidable conditions are congestive heart failure, electrolyte imbalance/dehydration, respiratory infection, sepsis, urinary tract or kidney infection, hypoglycemia or diabetic complications, anticoagulant complications, fractures and musculoskeletal injuries, acute delirium, adverse drug reactions, cellulitis/wound infections, pressure ulcers, and abnormal blood pressure.
• The two patient assessment–based, risk-adjusted measures of change in functional status were essentially unchanged between 2012 and 2018. The mobility measures are composites of the patients’ abilities in bed mobility, transfer, and ambulation, and they reflect the likelihood that a patient’s abilities will change, given his or her functional ability at admission. A facility admitting patients with worse prognoses will have lower expected rates of achieving these outcomes, and this difference will be reflected in the risk-adjusted rates.
A Data Book: Health care spending and the Medicare program, July 2020 105
Chart 8-9. Trends in the provision of home health care Percent change 2011–2018 2011 2018 Annual average Cumulative Number of users (in millions) 3.4 3.4 –0.3 –1.9 Share of FFS beneficiaries who used home health care 9.4% 8.8 –1.1 –7.2 Episodes (in millions) 6.8 6.3 –1.2 –8.2 Episodes per home health patient 2.0 1.9 –0.9 –6.4 Visits per home health episode 17.2 16.5 –0.6 –4.0 Visits per home health patient 34.2 30.8 –1.5 –10.2 Average payment per episode $2,917 $3,089 0.8 5.9 Note: FFS (fee-for-service). Yearly figures presented in the table are rounded, but the percent-change columns were calculated
using unrounded data. Average payment per episode excludes low-use episodes, those with fewer than five visits.
Source: MedPAC analysis of the home health standard analytic file from CMS.
• Between 2011 and 2018, episode volume declined by 8.2 percent and the number of users
dropped 1.9 percent.
• The number of visits per patient decreased between 2011 and 2018. This decline was a consequence of two other utilization declines in this period: a decline in average number of episodes per home health patient and a decline in the average number of visits per episode.
• The average payment per full episode was $3,089 in 2018, an increase of 5.9 percent relative to 2011. Throughout the 2011 to 2018 period, Medicare implemented a number of policies to reduce or slow the growth in home health payments. However, despite these reductions, the margins of free-standing home health agencies were over 15 percent in 2017 and 2018, indicating that payments remain well in excess of costs despite these policies (see Chart 8-11).
106 Post-acute care
Chart 8-10. Most home health episodes are not preceded by hospitalization or PAC stay
Number of episodes Percent change (in millions) 2011‒2018
2011 2018 Annual average Cumulative Episodes preceded by a hospitalization or PAC stay 2.2 2.1 >–0.1% –0.5% Episodes not preceded by a hospitalization or PAC stay 4.6 4.2 –1.4 –10.3 Share of episodes not preceded by a hospitalization or PAC stay 67% 66% –0.4 –2.7 Total 6.8 6.3 –1.1 –7.8
Note: PAC (post-acute care). “Episodes preceded by a hospitalization or PAC stay” refers to episodes that occurred less than 15 days after a stay in a hospital (including a long-term care hospital), skilled nursing facility, or inpatient rehabilitation
facility. “Episodes not preceded by a hospitalization or PAC stay” refers to episodes for which there was no hospitalization or PAC stay in the previous 15 days.
Source: 2018 home health standard analytic file, 2018 Medicare Provider and Analysis Review file, and 2018 skilled nursing facility standard analytic file.
• Most home health episodes are not preceded by a hospitalization or institutional PAC stay, and these episodes accounted for about two-thirds of PAC stays in 2011 through 2018. During this period, the number of home health episodes not preceded by a hospitalization or PAC stay declined 10.3 percent, while the number of episodes preceded by a hospitalization or PAC stay decreased 0.5 percent.
• The experience of the 2011 through 2018 period follows one that saw large growth in the number and share of episodes not preceded by a hospital or institutional PAC stay (data not shown). In 2001, episodes not preceded by a hospital or institutional PAC stay accounted for 53 percent of volume; by 2011 those episodes had increased to 67 percent of total episodes. Over the same period, the share of episodes preceded by a hospitalization or institutional PAC stay declined from 47 percent in 2001 to 33 percent in 2011 (data not shown). The shares of episode volume accounted for by these two categories have not changed substantially since 2011.
• Beneficiaries for whom the majority of home health episodes were preceded by a hospitalization or PAC stay had different characteristics from community-admitted beneficiaries (those who had no prior hospitalization or PAC) (data not shown). These beneficiaries were more likely to be dually eligible for Medicare and Medicaid, to have more home health episodes, and to have more episodes with a high share of home health aide services compared with other home health users coming from a hospitalization or other PAC stay (data not shown). Community-admitted users generally had slightly fewer chronic conditions, tended to be older, and were more likely to have dementia or Alzheimer’s disease (data not shown).
A Data Book: Health care spending and the Medicare program, July 2020 107
Chart 8-11. Medicare margins for freestanding home health agencies, 2017 and 2018
Share of agencies 2017 2018 2018 All 15.2% 15.3% 100% Geography Mostly urban 15.8 15.6 84 Mostly rural 13.4 13.8 16 Type of control For profit 16.4 16.8 89 Nonprofit 10.9 9.9 11 Volume quintile (lowest to highest) First 7.4 7.8 20 Second 9.8 9.3 20 Third 11.5 11.9 20 Fourth 13.6 13.9 20 Fifth 17.0 17.3 20
Note: Agencies are characterized as urban or rural based on the residence of the majority of their patients. Source: MedPAC analysis of 2017–2018 Medicare Cost Report files from CMS.
• In 2018, freestanding home health agencies (HHAs) (85 percent of all HHAs) had an
aggregate margin of 15.3 percent. HHAs that served mostly urban patients in 2018 had an aggregate margin of 15.6 percent; HHAs that served mostly rural patients had an aggregate margin of 13.8 percent. The 2018 margin is consistent with the historically high margins the home health industry has experienced since the prospective payment system (PPS) was implemented in 2000. The margins from 2001 to 2017 averaged 16.5 percent (data not shown), indicating that most agencies have been paid well in excess of their costs under the PPS.
• For-profit agencies in 2018 had an average margin of 16.8 percent, and nonprofit agencies had an average margin of 9.9 percent.
• Agencies with higher episode volumes had higher margins. The agencies in the lowest volume quintile in 2018 had an aggregate margin of 7.8 percent, while those in the highest quintile had an aggregate margin of 17.3 percent.
108 Post-acute care
Chart 8-12. Home health agencies’ assessment-based performance measures increased markedly from 2014 to 2018, while claims-based performance measures were largely unchanged
Measure 2014 2015 2016 2017 2018
Average share of an agency’s beneficiaries who:
Used emergency department care 12.0% 12.2% 12.1% 12.7% 12.8%
Had to be admitted to the hospital 15.4 15.5 16.2 15.4 15.4
Average share of a home health agency’s beneficiaries with improvements in:
Walking 61 63 69 74 77
Transferring 55 59 65 72 77
Note: All data pertain to fee-for-service beneficiaries only and are risk adjusted for differences in patient condition among home
health patients.
Source: MedPAC analysis of Medicare claims data and Outcome and Assessment Information Set data provided by the University
of Colorado.
• Quality measures for home health care draw on two sources, claims for payment submitted by home health agencies (HHAs) and patient assessment data collected by HHAs. In recent years, quality measures based on claims have indicated little change in quality, while measures based on patient assessment data have indicated improved quality. The claims-based rates of hospitalization and emergency department use have not changed significantly from 2014 to 2018, while the patient assessment–based functional improvement rates have improved substantially. From 2014 and 2018, average rates of beneficiaries with improvement in transferring improved from 55 percent to 77 percent. These divergent trends raise concerns about the objectivity of the patient assessment data and suggest that the functional measures of quality, such as walking and transferring, should be interpreted carefully.
• Medicare implemented a value-based purchasing program for home health agencies in nine states in 2018. In 2020, agencies in these states will receive bonuses or penalties of up to 6 percent, depending on their performance on 20 measures, including the functional and emergency department use measures listed above.
A Data Book: Health care spending and the Medicare program, July 2020 109
Chart 8-13. Number of FFS IRF cases increased in 2018 Average annual percent Percent change change
2010 2015 2017 2018 2010–2018 2017–2018 Number of IRF cases 365,095 393,475 396,294 408,038 1.4% 3.0% Cases per 10,000 101.3 103.4 102.7 105.7 0.5 2.9 FFS beneficiaries Payment per case $16,814 $18,527 $19,481 $20,124 2.3% 3.3 Average length of stay (in days) 13.1 12.7 12.7 12.7 –0.4 –0.6
Note: FFS (fee-for-service), IRF (inpatient rehabilitation facility). Numbers of cases reflect Medicare FFS utilization only. The
number of cases presented differs from past reports due to a change in methodology. Yearly figures presented in the
table are rounded, but the percent-change columns were calculated using unrounded data.
Source: MedPAC analysis of Medicare Provider Analysis and Review data from CMS.
• After a period of relative stability from 2015 to 2017, the number of Medicare FFS cases increased 3.0 percent, growing to about 408,000 cases in 2018.
• In 2018, the number of IRF cases per 10,000 FFS beneficiaries grew to 105.7, up 2.9 percent from the previous year. Relatively few Medicare beneficiaries use IRF services because, to qualify for Medicare coverage, IRF patients must be able to tolerate and benefit from rehabilitation therapy that is intensive, which is usually interpreted to mean at least three hours of therapy a day for at least five days a week. Yet, compared with all Medicare beneficiaries, those admitted to IRFs in 2018 were disproportionately over age 85 (data not shown).
• With the increase in the number of IRF cases per FFS beneficiary, FFS Medicare’s share of IRF discharges rose slightly, to 59 percent of total discharges (data not shown).
• The average length of stay in an IRF has held steady since 2015.
110 Post-acute care
Chart 8-14. Most common types of FFS inpatient rehabilitation facility cases, 2018
Type of case Share of cases
Stroke 20.0% Other neurological conditions 14.7 Brain injury 10.8 Debility 11.6 Fracture of the lower extremity 10.3 Other orthopedic conditions 7.9 Cardiac conditions 5.9 Spinal cord injury 4.9 Major joint replacement of lower extremity 4.1 All other 9.7
Note: FFS (fee-for-service). “Other neurological conditions” includes multiple sclerosis, Parkinson’s disease, polyneuropathy,
and neuromuscular disorders. “Fracture of the lower extremity” includes hip, pelvis, and femur fractures. Patients with debility have generalized deconditioning not attributable to other conditions. “Other orthopedic conditions” excludes fractures of the hip, pelvis, and femur and hip and knee replacements. “All other” includes conditions such as
amputations, arthritis, and pain syndrome. All Medicare FFS inpatient rehabilitation facility (IRF) cases with valid patient assessment information were included in this analysis.
Source: MedPAC analysis of Inpatient Rehabilitation Facility–Patient Assessment Instrument data from CMS.
• In 2018, the most frequently occurring case type among FFS beneficiaries admitted to IRFs was stroke, which accounted for 20.0 percent of Medicare FFS cases.
• Between 2017 and 2018, we observed disproportionate growth in the number of cases with debility (data not shown). Between 2017 and 2018, the share of these cases rose from 10.7 percent to 11.6 percent of FFS IRF cases.
• The distribution of case types differs by type of IRF (data not shown). For example, in 2018, only 16 percent of cases in freestanding for-profit IRFs were admitted for rehabilitation following a stroke, compared with 26 percent of cases in hospital-based nonprofit IRFs. Likewise, 20 percent of cases in freestanding for-profit IRFs were admitted with other neurological conditions, twice the share admitted to hospital-based nonprofit IRFs. Cases with other orthopedic conditions also made up a higher share of cases in freestanding for-profit facilities than in all other IRFs. By contrast, the share of cases with brain injury or debility was similar across IRF types (data not shown).
A Data Book: Health care spending and the Medicare program, July 2020 111
Chart 8-15. Inpatient rehabilitation facilities’ Medicare margins by type of facility, 2010–2018 2010 2012 2014 2015 2016 2017 2018 All IRFs 8.6% 11.2% 12.2% 13.9% 13.3% 13.9% 14.7% Hospital based –0.5 0.6 0.7 2.1 0.8 1.5 2.5 Freestanding 21.4 23.9 25.2 26.6 25.8 25.6 25.4 Urban 9.0 11.5 12.6 14.3 13.6 14.2 15.0 Rural 4.7 6.6 6.4 8.6 9.1 8.2 9.8 Nonprofit 2.1 2.0 1.7 3.4 1.5 2.1 2.4 For profit 19.6 23.0 23.9 25.1 24.5 24.1 24.6 Note: IRF (inpatient rehabilitation facility).
Source: MedPAC analysis of cost report data from CMS.
• In 2018, the aggregate IRF Medicare margin increased to 14.7 percent.
• Margins varied by ownership, with for-profit IRFs having substantially higher margins. At the same time, Medicare margins in freestanding IRFs far exceeded those of hospital-based
facilities. Nevertheless, one-quarter of hospital-based IRFs had Medicare margins greater than 13 percent (data not shown), indicating that many hospitals can manage their IRF units profitably. Further, despite comparatively low average margins in hospital-based IRFs, evidence suggests that these units make a positive financial contribution to their parent hospitals. For example, aggregate inpatient Medicare margins for hospitals are consistently higher for hospitals with IRF units versus hospitals without IRF units (1.4 percentage points higher in 2018). Aggregate overall Medicare margins for hospitals with IRF units were 2.4 percentage points higher for 2018 (data not shown).
• Higher unit costs are a major driver of low margins in both hospital-based and nonprofit IRFs. However, in an analysis of data from 2013, the Commission found that the mix of case types in IRFs was also correlated with profitability. IRFs with the highest margins had a higher share of neurological cases and a lower share of stroke cases. Further, we observed differences in the types of stroke and neurological cases admitted to high- and low-margin IRFs. Stroke cases in the highest margin IRFs were much less likely to have paralysis than were stroke cases in the lowest margin IRFs. Neurological cases in the highest margin IRFs were much more likely to be
neuromuscular disorders (such as amyotrophic lateral sclerosis or muscular dystrophy) than were neurological cases in the lowest margin IRFs (data not shown).
• In an analysis of data from 2013, the Commission found that high-margin IRFs had patients who were, on average, less severely ill in the preceding stay in an acute care hospital than patients admitted to low-margin IRFs. Once admitted to and assessed by the IRF, however, the average patient profile changed, with patients treated in high-margin IRFs appearing to be more disabled than those in low-margin IRFs. This finding suggests the possibility that assessment and coding practices may contribute to greater revenues in some IRFs (data not shown).
112 Post-acute care
Chart 8-16. Low standardized costs led to high margins for both hospital-based and freestanding IRFs, 2018
Characteristic Lowest cost quartile Highest cost quartile
Median cost per discharge All $11,583 $20,257 Hospital based 12,216 20,278 Freestanding 11,194 20,001 Median Medicare margin
All 28.6% −19.9%
Hospital based 23.7 −20.4
Freestanding 31.4 −17.3 Median Number of beds 44 18 Occupancy rate 74% 54% Share of facilities in the quartile that are: Hospital based 36% 94% Freestanding 64 6 Nonprofit 25 64 For profit 70 18 Government 5 17 Urban 94 72 Rural 6 28
Note: IRF (inpatient rehabilitation facility). Cost per discharge is standardized for differences in wages across geographic areas, differences in case mix across providers, and differences across providers in the prevalence of high-cost outliers, short-
stay outliers, and transfer cases. Totals may not sum to 100 percent due to rounding. Source: MedPAC analysis of Medicare cost report and Medicare Provider Analysis and Review data from CMS.
• IRFs with the lowest standardized costs (those in the lowest cost quartile) had a median standardized cost per discharge that was 43 percent less than that of the IRFs with the highest standardized costs (those in the highest cost quartile).
• IRFs with the lowest costs tended to be larger: The median number of beds was 44 compared with 18 in the highest cost quartile. In addition, IRFs with the lowest costs had a higher median occupancy rate (74 percent vs. 54 percent, respectively). These results suggest that low-cost IRFs benefit from economies of scale.
• Low-cost IRFs were disproportionately freestanding and for profit. Still, 36 percent of IRFs in the lowest cost quartile were hospital based and 25 percent were nonprofit. By contrast, in the highest cost quartile, 94 percent were hospital based and 64 percent were nonprofit.
A Data Book: Health care spending and the Medicare program, July 2020 113
Chart 8-17. Risk-adjusted quality indicators for IRFs held steady or improved slightly from 2012 to 2018
Measure 2012 2014 2015 2016 2017 2018
Potentially avoidable rehospitalizations during IRF stay 2.8% 2.7% 2.6% 2.7% 2.7% 2.6%
Potentially avoidable rehospitalizations during 30 days after discharge from IRF 5.0 4.8 4.4 4.8 4.8 4.8
Discharged to the community 74.4 75.3 75.1 76.0 76.0 76.4
Discharged to a SNF 6.7 6.9 6.9 6.7 6.7 6.6
Note: IRF (inpatient rehabilitation facility), SNF (skilled nursing facility). High rates of rehospitalization and discharge to a SNF indicate worse quality. High rates of discharge to the community indicate better quality. Rates are the average of the
facility rates and are calculated for all facilities with 25 or more stays.
Source: Analysis of Medicare claims data and Inpatient Rehabilitation Facility‒Patient Assessment Instrument data from CMS.
• Between 2012 and 2018, the national average rate of risk-adjusted potentially avoidable rehospitalizations during IRF stays declined from 2.8 percent to 2.6 percent (lower rates are better). The national average rate of risk-adjusted potentially avoidable rehospitalizations within 30 days after discharge from an IRF declined from 5.0 percent to 4.4 percent in 2015, then rose to 4.8 percent in 2016 and held steady.
• The rehospitalization rates count only stays readmitted to a hospital with the principal diagnosis of a potentially avoidable condition. The potentially avoidable rehospitalizations we measure are respiratory-related illness (pneumonia, influenza, bronchitis, chronic obstructive pulmonary disease, and asthma); sepsis; congestive heart failure; fractures or fall with a major injury; urinary tract or kidney infection; blood pressure management; electrolyte imbalance; anticoagulant therapy complications; diabetes-related complications; cellulitis or wound infection; pressure ulcer; medication error or adverse drug reaction; and delirium.
• Between 2012 and 2018, the national average for the risk-adjusted community discharge rate increased from 74.4 percent to 76.4 percent (higher rates are better). Our measure of community discharge does not give IRFs credit for discharging a Medicare beneficiary to the community if the beneficiary is subsequently admitted to an acute care hospital within 30 days of the IRF discharge.
• Between 2012 and 2018, the national risk-adjusted rate of discharge to a SNF improved slightly (was lower). Between 2012 and 2014, the national risk-adjusted rate of discharge to a SNF increased from 6.7 percent to 6.9 percent, but subsequently declined to 6.7 percent in 2016, then again in 2018 to 6.6 percent (lower rates are better).
114 Post-acute care
Chart 8-18. The top 25 MS–LTC–DRGs accounted for almost 70 percent of LTCH discharges in 2018
MS–LTC Share –DRG Description Discharges of cases 189 Pulmonary edema and respiratory failure 18,761 18.3% 207 Respiratory system diagnosis with ventilator support 96+ hours 12,691 12.4 871 Septicemia without ventilator support 96+ hours with MCC 6,154 6.0 166 Other respiratory system OR procedures with MCC 2,636 2.6 208 Respiratory system diagnosis with ventilator support <96 hours 2,616 2.6 949 Aftercare with CC/MCC 2,128 2.1 592 Skin ulcers with MCC 2,066 2.0 177 Respiratory infections and inflammations with MCC 2,038 2.0 981 Extensive OR procedure unrelated to principal diagnosis with MCC 1,938 1.9 539 Osteomyelitis with MCC 1,798 1.8 682 Renal failure with MCC 1,708 1.7 291 Heart failure and shock with MCC 1,645 1.6 4 Tracheostomy with ventilator support 96+ hours or primary diagnosis 1,542 1.5 except face, mouth, and neck without major OR 559 Aftercare, musculoskeletal system and connective tissue with MCC 1,459 1.4 314 Other circulatory system diagnoses with MCC 1,305 1.3 919 Complications of treatment with MCC 1,262 1.2 862 Postoperative and post-traumatic infections with MCC 1,249 1.2 853 Infectious and parasitic diseases with OR procedure with MCC 1,239 1.2 870 Septicemia with ventilator support 96+ hours 1,187 1.2 570 Skin debridement with MCC 1,108 1.1 193 Simple pneumonia and pleurisy w MCC 1,013 1.0 190 Chronic obstructive pulmonary disease with MCC 996 1.0 638 Diabetes with CC 986 1.0 560 Aftercare, musculoskeletal system and connective tissue w CC 968 1.0 689 Kidney and urinary tract infections with MCC 941 0.9 Top 25 MS–LTC–DRGs 71,434 70.0
Total 102,288 100.0 Note: MS–LTC–DRG (Medicare severity long-term care diagnosis related group), LTCH (long-term care hospital), MCC (major
complication or comorbidity), OR (operating room), CC (complication or comorbidity). MS–LTC–DRGs are the case-mix system for LTCHs.
Source: MedPAC analysis of Medicare Provider Analysis and Review data from CMS.
• Cases in LTCHs are concentrated in a relatively small number of MS–LTC–DRGs. In 2018,
the top 25 MS–LTC–DRGs accounted for 70 percent of LTCH Medicare cases.
• Consistent with 2016 and 2017, the two most frequent diagnoses in LTCHs in 2018 were pulmonary edema and respiratory failure and a respiratory system diagnosis with ventilator support of more than 96 hours.
• Over 40 percent of all LTCH cases were respiratory conditions. Nonprofit LTCHs care for a higher share of beneficiaries with diagnoses of pulmonary edema and respiratory failure and a respiratory system diagnosis with ventilator support of more than 96 hours than for-profit LTCHs (data not shown).
A Data Book: Health care spending and the Medicare program, July 2020 115
Chart 8-19. The number of Medicare LTCH cases and users decreased by over 11 percent between 2017 and
2018 Average annual change
2014– 2016– 2017– 2014 2015 2016 2017 2018 2016 2017 2018 Cases 133,984 131,129 125,586 116,424 102,288 –3.2% –7.3% –12.1% Cases per 10,000 FFS beneficiaries 35.4 34.4 32.5 30.1 26.5 –4.2 –7.3 –11.9 Payment per case $40,015 $40,719 $40,656 $38,253 $40,105 0.8 –5.9 4.8 Length of stay (in days) 26.3 26.6 26.8 26.3 26.6 1.0 –2.2 1.2 Users 118,288 116,088 111,171 103,322 91,754 –3.1 –7.1 –11.2
Note: LTCH (long-term care hospital), FFS (fee-for-service). Yearly figures presented in the table are rounded, but the average annual changes were calculated using unrounded data.
Source: MedPAC analysis of Medicare Provider Analysis and Review data from CMS.
• The Pathway for SGR Reform Act of 2013 created a “dual payment-rate structure” for LTCHs where, beginning in fiscal year 2016, only certain LTCH cases continue to qualify for the standard LTCH prospective payment system (PPS) rate, while cases that do not meet a set of criteria are paid a lower “site-neutral” rate (site neutral is defined as the lower of Medicare’s inpatient hospital PPS rate or 100 percent of the cost of the case).
• Controlling for the number of FFS beneficiaries, the number of LTCH cases declined by about 4 percent annually between 2014 and 2016. The number of cases declined more rapidly following the implementation of the dual payment-rate structure. From 2016 to 2017, the number of LTCH cases declined by over 7 percent, and from 2017 to 2018, the number of LTCH cases declined by another 12 percent.
• Reflecting the decline in the number of Medicare cases, the number of beneficiaries who had LTCH stays (“users”) also decreased by 11.2 percent from 2017 to 2018.
• Reductions in payment per case from 2015 through 2017 reflect a lower payment rate for cases that did not meet the criteria after the implementation of the dual payment-rate structure. From 2017 to 2018, on a payment per case basis, the increase in the share of cases that qualified for the standard LTCH PPS rate offset the payment reduction for cases paid the “site-neutral” rate.
116 Post-acute care
Chart 8-20. The aggregate LTCH Medicare margin increased in 2018
Type of LTCH
Share of discharges
in 2018
Medicare margin
2014 2015 2016 2017 2018
All 100% 5.2% 4.7% 3.9% −2.2% −0.5%
Urban 95 5.2 4.7* 4.0 −1.9 −0.2
Rural 5 5.1 3.5* −0.2 −13.6 −9.5
Nonprofit 14 −2.2 −5.9 −5.7 −13.0 −11.7
For profit 84 7.0 6.5 5.5 −0.3 1.3
Note: LTCH (long-term care hospital). *CMS adopted new core-based statistical area codes for LTCHs beginning in fiscal year 2015; this change reclassified
several facilities as urban that had previously been classified as rural, and therefore the margin across categories of urban
and rural facilities between 2014 and 2015 should not be compared. Source: MedPAC analysis of cost report data from CMS.
• After peaking in 2012, the aggregate LTCH margin began to fall in 2013, primarily due to policy changes that reduced payments, including the start of a three-year phase-in of a downward adjustment for budget neutrality and the effect of sequestration that began on April 1, 2013. Margins steadily declined between 2012 and 2018 (early years not shown).
• In fiscal year 2016, CMS began implementing a “dual payment-rate structure” where certain LTCH cases not meeting a set of criteria specified in law are paid a lower “site-neutral” rate (site neutral is defined as the lower of Medicare’s inpatient hospital PPS rate or 100 percent
of the cost of the case). As a result, the aggregate Medicare margin fell to −2.2 percent in 2017. Increases in the aggregate share of cases meeting the criteria resulted in the aggregate Medicare margin increasing to –0.5 percent in 2018.
• Financial performance in 2018 varied across LTCHs. The aggregate Medicare margin for for-profit LTCHs (which accounted for 84 percent of all Medicare discharges from LTCHs) decreased from 6.5 percent in 2015 to 1.3 percent in 2018. The aggregate margin for nonprofit LTCHs decreased from –5.9 percent in 2015 to –11.7 percent in 2018.
A Data Book: Health care spending and the Medicare program, July 2020 117
Chart 8-21. The share of LTCH cases meeting the criteria for the standard LTCH PPS rate increased from 2017 to 2018
Percent change
Cases meeting the
criteria 2015 2016 2017 2018
2015–
2017
2017–
2018
Cases 72,429 72,318 74,666 71,916 1.5% –3.7%
Share of all LTCH
cases 55% 58% 64% 70%
Cases per 10,000 FFS
beneficiaries 19.0 18.7 19.3 18.6 0.7 –3.4
Payment per case $46,217 $46,223 $46,127 $46,789 –0.1 1.4
Spending (in billions) $3.3 $3.3 $3.4 $3.4 1.4 –2.3
Length of stay (in days) 28.5 27.9 27.9 28.0 –1.0 0.4
Aggregate Medicare
margin 6.8% 6.3% 5.8% 5.8% N/A N/A
Note: LTCH (long-term care hospital), PPS (prospective payment system), FFS (fee-for-service), N/A (not applicable). Yearly
figures presented in the table are rounded, but the percent changes were calculated using unrounded data. Source: MedPAC analysis of cost report data from CMS.
• The Pathway for SGR Reform Act of 2013 created a “dual payment-rate structure” for LTCHs where, beginning in fiscal year 2016, only certain LTCH cases continue to qualify for the standard LTCH PPS rate, while cases that do not meet a set of criteria are paid a lower “site-neutral” rate (site neutral is defined as the lower of Medicare’s inpatient hospital PPS rate or 100 percent of the cost of the case).
• Controlling for the number of FFS beneficiaries, the number of cases meeting the criteria to qualify for the standard LTCH PPS rate decreased by 3.4 percent in 2018, in contrast to the 11.9 percent reduction in all LTCH cases per 10,000 FFS beneficiaries (see Chart 8-19).
• After decreasing from 28.5 days in 2015 to 27.9 days in 2016, the average length of stay for cases meeting the criteria to qualify for the standard LTCH PPS rate has remained relatively stable since 2017.
• The aggregate Medicare margin for cases meeting the criteria to qualify for the standard LTCH PPS rate was 5.8 percent in 2018. Because cases that meet the criteria are generally more profitable under the dual payment-rate structure than those that do not, we expect stronger financial performance under Medicare for LTCHs that treat higher shares of these cases.
9
A Data Book: Health care spending and the Medicare program, July 2020 121
Chart 9-1. Enrollment in MA plans, 2003–2020
Note: MA (Medicare Advantage).
Source: CMS Medicare managed care contract reports and monthly summary reports, February 2003–2020.
• Historically, the Commission has used information on “Medicare eligibles” as the denominator in calculating the share of Medicare beneficiaries enrolled in Medicare Advantage (MA). “Medicare eligibles” include people previously, but no longer, covered by Medicare and people within 5 months of their 65th birthday. We now have data that allows us to calculate the share of MA enrollment as a share of Medicare beneficiaries with either Part A or Part B coverage and thus can calculate a more accurate MA enrollment percentage. The percentages published here supersede all prior estimates by the Commission of the share of Medicare beneficiaries enrolled in MA.
• Enrollment in MA plans that are paid on an at-risk capitated basis reached 24.0 million enrollees in February 2020. MA enrollment represents 39 percent of all 62.2 million Medicare beneficiaries (and 42 percent of all 56.5 million beneficiaries enrolled in both Part A and Part B). Medicare Advantage and other private plans account for 40 percent of all Medicare beneficiaries. (Other private plans consist of private fee-for-service plans, cost plans, Medicare medical savings account plans, plans under the Program of All-Inclusive Care for the Elderly (PACE), and Medicare–Medicaid plans participating in CMS’s financial alignment demonstration.)
• MA enrollment has grown steadily since 2003, increasing more than fivefold. The Medicare program paid MA plans about $274 billion in 2019 to cover Part A and Part B services for MA enrollees (data not shown).
4.6 4.7 4.9
6.98.1
9.410.5 11.0
11.712.8
14.115.4
16.417.2
18.520.0
21.9
24.0
0
5
10
15
20
25
30
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Ben
efi
cia
ries (
in m
illi
on
s)
122 Medicare Advantage
Chart 9-2. MA plans available to almost all Medicare beneficiaries, 2012–2020
Share of Medicare beneficiaries living in counties with plans available
CCPs
HMO or Any Average plan local PPO Regional Any MA offerings per (local CCP) PPO CCP PFFS plan beneficiary 2012 93% 76% 99% 60% 100% 19 2013 95 71 99 59 100 19 2014 95 71 99 53 100 18 2015 95 70 98 47 99 17 2016 96 73 99 47 99 18 2017 95 74 98 45 99 18 2018 96 74 98 41 99 20 2019 97 74 98 38 99 23 2020 98 73 99 36 99 27 Note: MA (Medicare Advantage), CCP (coordinated care plan), HMO (health maintenance organization), PPO (preferred
provider organization), PFFS (private fee-for-service). These data do not include plans that have restricted enrollment or are not paid based on the MA plan bidding process (special needs plans, cost plans, employer-only plans, and certain
demonstration plans). Source: MedPAC analysis of plan bid data from CMS, 2012–2020.
• There are four types of MA plans, three of which are CCPs. Local CCPs include HMOs and local PPOs, which have comprehensive provider networks and limit or discourage use of out-of-network providers. Local CCPs may choose which individual counties to serve. Regional PPOs cover entire state-based regions and have networks that may be looser than those of local PPOs. These CCPs accounted for 97 percent of Medicare private plan enrollees as of February 2020 (data not shown). Since 2011, PFFS plans are required to have networks in areas with two or more CCPs. In other areas, PFFS plans are not required to have networks, and enrollees are free to use any Medicare provider.
• Local CCPs are available to 98 percent of Medicare beneficiaries in 2020, and regional PPOs are available to 73 percent of beneficiaries. Since 2006, almost all Medicare beneficiaries have had MA plans available (data not shown); 99 percent have an MA plan available in 2020.
• The number of plans from which beneficiaries may choose in 2020 is higher than at any time during the years examined. In 2020, beneficiaries can choose from an average of 27 plans operating in their counties.
A Data Book: Health care spending and the Medicare program, July 2020 123
Chart 9-3. Average monthly rebate dollars, by plan type, 2015–2020
Note: HMO (health maintenance organization), PPO (preferred provider organization), PFFS (private fee-for-service), MA
(Medicare Advantage). Employer group waiver and special needs plans are excluded.
Source: MedPAC analysis of bid and plan finder data from CMS.
• Perhaps the best summary measure of plan benefit value is the average rebate, which plans receive to provide additional benefits. Plans are awarded rebates for bidding under their benchmarks. The rebates must be returned to the plan members in the form of extra benefits. The extra benefits may be lower cost sharing, supplemental benefits, or lower premiums. The average rebate for all non-employer, non–special needs plans rose to a high of $122 per month per beneficiary for 2020.
• HMOs have had, by far, the highest rebates because they tend to bid lower than other types of plans. Average rebates for HMOs have risen sharply over the past few years and are at a high of $141 per month per beneficiary for 2020.
• For both local and regional PPOs, the rebates rose sharply after 2016. Rebates for local PPOs have tripled since 2015.
• Rebates for PFFS plans have been relatively stable since 2015.
95
29
18
24
76
101
34
17
26
81
105
49
16
40
89
111
59
15
45
95
125
71
18
46
107
141
86
17
49
122
0
20
40
60
80
100
120
140
160
HMO Local PPO PFFS Regional PPO Any MA plan
Avera
ge m
on
thly
reb
ate
do
llars
per
ben
efi
cia
ry 2015 2016 2017 2018 2019 2020
124 Medicare Advantage
Chart 9-4. Changes in enrollment vary among major plan types Total enrollees (in thousands) Percent change Plan type 2016 2017 2018 2019 2020 2019–2020 Local CCPs 15,588 16,920 18,463 20,502 22,704 11% Regional PPOs 1,315 1,353 1,327 1,255 1,170 –7 PFFS 238 190 154 118 87 –26 Note: CCP (coordinated care plan), PPO (preferred provider organization), PFFS (private fee-for-service). Local CCPs include
HMOs and local PPOs.
Source: CMS health plan monthly summary reports, February 2016–2020.
• Enrollment in local CCPs grew by 11 percent over the past year. Enrollment in regional PPOs declined by 7 percent, and enrollment in PFFS plans dropped by 26 percent. Combined enrollment in the three types of plans grew by 10 percent from February 2019 to February 2020 (data not shown).
A Data Book: Health care spending and the Medicare program, July 2020 125
Chart 9-5. MA and cost plan enrollment by state and type of plan, 2020
All Medicare beneficiaries Distribution (in percent) of enrollees by plan type
State or territory (in thousands) HMO Local PPO Regional PPO PFFS Cost Total
U.S. total 62,171 24% 13% 2% 0% 0% 39% Alabama 1,054 21 23 1 0 0 45 Alaska 102 0 1 0 0 0 1
Arizona 1,342 35 7 1 0 0 43 Arkansas 641 16 5 7 1 0 29 California 6,350 40 3 0 0 0 43
Colorado 926 31 11 0 0 1 43 Connecticut 686 21 22 1 0 0 44 Delaware 211 8 10 0 0 0 18
Florida 4,608 32 12 5 0 0 49 Georgia 1,748 13 21 9 0 0 42 Hawaii 277 17 28 2 0 0 47
Idaho 340 22 14 0 0 0 37 Illinois 2,256 12 15 0 0 0 27 Indiana 1,271 14 19 2 0 0 35
Iowa 631 9 14 0 0 2 25 Kansas 540 8 12 0 1 0 21 Kentucky 935 14 20 3 0 1 38
Louisiana 880 35 6 1 0 0 41 Maine 342 22 16 1 0 0 39 Maryland 1,048 8 4 0 0 0 13
Massachusetts 1,344 16 7 1 0 0 24 Michigan 2,084 17 28 0 0 0 45 Minnesota 1,033 15 27 0 0 6 48
Mississippi 607 13 7 3 0 0 23 Missouri 1,238 24 12 3 0 0 40 Montana 233 9 11 0 0 0 20
Nebraska 350 12 6 0 1 1 20 Nevada 538 34 6 0 0 0 40 New Hampshire 303 10 9 2 0 0 21
New Jersey 1,630 15 17 0 0 0 32 New Mexico 427 21 17 0 0 0 39 New York 3,663 28 11 3 0 0 43
North Carolina 2,006 17 20 3 0 0 40 North Dakota 133 0 4 0 0 15 20 Ohio 2,359 25 17 1 0 0 43
Oklahoma 745 13 11 1 0 0 25 Oregon 874 31 17 0 0 0 47 Pennsylvania 2,756 27 17 0 0 0 44
Puerto Rico 752 78 2 0 0 0 80 Rhode Island 223 36 5 1 0 0 42 South Carolina 1,088 10 11 10 0 0 31
South Dakota 178 1 8 0 0 13 23 Tennessee 1,368 28 14 1 0 0 43 Texas 4,219 23 14 4 0 0 42
Utah 406 31 9 0 0 0 40 Vermont 149 5 5 4 0 0 14 Virgin Islands 20 0 0 0 0 0 1
Virginia 1,527 14 6 2 1 1 25 Washington 1,380 31 5 0 0 0 36 Washington, DC 94 8 13 0 0 0 22
West Virginia 441 4 29 1 1 4 38 Wisconsin 1,186 25 16 1 0 4 46 Wyoming 112 0 2 0 2 1 4
Note: MA (Medicare Advantage), HMO (health maintenance organization), PPO (preferred provider organization), PFFS (private fee-for-service). Cost plans are not MA plans; they submit cost reports rather than bids to CMS. U.S. total includes beneficiaries in all outlying areas. Component percentages may not sum to totals due to rounding. In contrast with prior years, we report MA enrollment as a share of current Medicare beneficiaries. In prior years, we reported MA enrollment as
a percentage of total Medicare eligibles, which included individuals who were (1) alive and ever enrolled but no longer in Medicare and (2) enrolled in Medicare with a future effective date.
Source: CMS enrollment and population data February 2020.
126 Medicare Advantage
Chart 9-6. MA plan benchmarks, bids, and Medicare program payments relative to FFS spending, 2020 All plans HMOs Local PPOs Regional PPOs PFFS Benchmarks/FFS 107% 107% 109% 105% 106% Bids/FFS 88 87 94 91 105 Payments/FFS 100 100 104 97 105
Note: MA (Medicare Advantage), FFS (fee-for-service), HMO (health maintenance organization), PPO (preferred provider
organization), PFFS (private fee-for-service). Data exclude employer plans, which do not submit plan bids but receive payment based on the bids and benchmarks of nonemployer plans.
Source: MedPAC analysis of plan bid data from CMS October 2019.
• Since 2006, plan bids have partly determined the Medicare payments that plans receive. Plans bid to offer Part A and Part B coverage to Medicare beneficiaries (Part D coverage is bid separately). The bid includes plan administrative cost and profit. CMS bases the Medicare payment for a private plan on the relationship between its bid and its applicable benchmark.
• The benchmark is an administratively determined bidding target. Benchmarks for each county are set by means of a statutory formula based on percentages (ranging from 95 percent to 115 percent) of each county’s per capita Medicare FFS spending. Plans with quality ratings of 4 or more stars may have their benchmarks raised by 10 percent in some counties.
• If a plan’s bid is above the benchmark, then the plan receives the benchmark as payment from Medicare, and enrollees have to pay an additional premium that equals the difference. If a plan’s bid is below the benchmark, the plan receives its bid plus a “rebate,” defined by law as a percentage of the difference between the plan’s bid and its benchmark. The percentage is based on the plan’s quality rating, and it ranges from 50 percent to 70 percent. The plan must then return the rebate to its enrollees in the form of lower cost sharing, supplemental benefits, or lower premiums.
• We estimate that MA benchmarks average 107 percent of FFS spending when weighted by MA enrollment. The ratio varies by plan type because different types of plans tend to draw enrollment from different types of geographical areas.
• Plans’ enrollment-weighted bids (excluding employer plans, which no longer submit bids) average 88 percent of FFS spending in 2020. On average, each coordinated care plan type (HMO, local PPO, regional PPO) has demonstrated the ability to provide the same services for less than FFS in the areas where they bid.
• We project that 2020 MA payments will be 100 percent of FFS spending. This figure does not include employer plans and does not account for risk-coding differences between FFS and MA plans that have not been resolved through the coding intensity factor. We estimate that coding differences add 2 percentage points to 3 percentage points to payments relative to FFS.
• The ratio of payments relative to FFS spending varies by the type of MA plan. HMO and regional PPO payments are estimated to be 100 and 97 percent of FFS, respectively, while payments to local PPOs and PFFS plans average 104 percent and 105 percent of FFS, respectively.
A Data Book: Health care spending and the Medicare program, July 2020 127
Chart 9-7. Enrollment in employer group MA plans, 2007–2020
Note: MA (Medicare Advantage). Enrollment numbers are as of November for 2007 and February for 2008 through 2020.
Source: CMS enrollment data.
• While most MA plans are available to any Medicare beneficiary residing in a given area, some MA plans are available only to retirees whose Medicare coverage is supplemented by their former employer or union. These plans are called employer group plans. Such plans are usually offered through insurers and are marketed to groups formed by employers or unions rather than to individual beneficiaries.
• As of February 2020, about 4.7 million enrollees were in employer group plans, or about 20 percent of all MA enrollees. Employer plan enrollment grew by 5 percent from 2019 and has doubled since 2012.
1.301.56
1.83 1.92 2.082.32 2.53
2.96 3.14 3.163.68
4.134.51 4.73
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
En
roll
men
t (i
n m
illi
on
s)
128 Medicare Advantage
Chart 9-8. Number of special needs plan enrollees, 2011–2020
Source: CMS special needs plans comprehensive reports, April 2011–2020.
• The Congress created special needs plans (SNPs) as a new MA plan type in the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 to provide a common framework for the existing plans serving special needs beneficiaries and to expand beneficiaries’ access to and choice among MA plans.
• SNPs were originally authorized for five years, but SNP authority was extended several times. The Bipartisan Budget Act of 2018 made SNPs permanent.
• CMS approves three types of SNPs: dual-eligible SNPs enroll only beneficiaries dually entitled to Medicare and Medicaid, chronic condition SNPs enroll only beneficiaries who have certain chronic or disabling conditions, and institutional SNPs enroll only beneficiaries who reside in institutions or are nursing-home certified.
• Enrollment in dual-eligible SNPs has grown continuously and exceeds 2.9 million in 2020, doubling since 2013.
• Enrollment in chronic condition SNPs has fluctuated as plan requirements have changed but has generally risen annually since 2011.
• Enrollment in institutional SNPs has risen annually since 2015.
1,0691,188
1,380
1,5761,678
1,790
1,959
2,210
2,532
2,929
170 201 266 288 313 332 335 349 349 375
80 47 49 50 50 58 63 74 87 104
0
500
1,000
1,500
2,000
2,500
3,000
3,500
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Nu
mb
er
of
sp
ecia
l n
eed
s p
lan
en
rollees
(in
th
ou
san
ds)
Dual Chronic Institutional
A Data Book: Health care spending and the Medicare program, July 2020 129
Chart 9-9. Number of SNPs and SNP enrollment rose from 2019 to 2020
Note: SNP (special needs plan).
Source: CMS special needs plans comprehensive reports, April 2019 and 2020.
• The number of SNPs increased by 19 percent from April 2019 to April 2020. Dual-eligible SNPs increased by 15 percent, institutional SNPs increased by 20 percent, and the number of chronic condition SNPs increased by 33 percent.
• In 2020, most SNPs (63 percent) are for dual-eligible beneficiaries, while 17 percent are for beneficiaries who reside in institutions (or reside in the community but have a similar level of need), and 20 percent are for beneficiaries with chronic conditions.
• From April 2019 to April 2020, the number of SNP enrollees increased by 15 percent. Enrollment in SNPs for dual eligibles grew by 16 percent, enrollment in SNPs for institutionalized beneficiaries grew by 19 percent, and enrollment in SNPs for chronic conditions grew by 7 percent. Enrollment in all SNPs has grown from 0.9 million in May 2007 (not shown) to 3.4 million in April 2020.
• The availability of SNPs varies by type of special needs population served (data not shown). In 2020, 90 percent of beneficiaries reside in areas where SNPs serve dual-eligible beneficiaries (up from 89 percent in 2019), 67 percent live where SNPs serve institutionalized beneficiaries (up from 63 percent in 2019), and 52 percent live where SNPs serve beneficiaries with chronic conditions (up from 47 percent in 2019).
480551
125
150129
171
0
100
200
300
400
500
600
700
800
900
1000
April 2019 April 2020
Nu
mb
er
of
SN
Ps
Chronic or disabling condition
Institutional
Dual eligible
2,5322,929
87
104349
375
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
April 2019 April 2020S
NP
en
roll
men
t (i
n t
ho
usan
ds)
Chronic or disabling condition
Institutional
Dual eligible
130 Medicare Advantage
Chart 9-10. The share of Medicare beneficiaries in private plans does not differ substantially in medically underserved areas compared with other areas, but is somewhat lower in rural areas, 2020
Medicare population (in millions)
As percent of Medicare population
Percent of category in MA
and other private plans
All beneficiaries 61.7 100% 40% County’s medically underserved area designation Partial county 38.8 63 41 Entire county 10.9 18 38 No medically underserved areas 12.0 19 37 Urban influence code designation Metropolitan 44.5 72 43 Rural: Micropolitan 7.9 13 33 Rural: Adjacent to metropolitan 8.5 14 32 Rural: Not adjacent to metropolitan 0.7 1 23
Note: MA (Medicare Advantage). Beneficiaries in the Virgin Islands, Guam, American Samoa, and the Northern Mariana Islands or in foreign areas are excluded. MA plans consist of HMOs, local preferred provider organizations (PPOs), regional PPOs, private fee-for-service plans, and Medical Savings Account plans. Private plans include 0.7 million beneficiaries in
non-MA private plans, which consist of Medicare–Medicaid plans, Program of All-Inclusive Care for the Elderly plans, and cost plans. Medically underserved areas (MUAs) are designated by the Health Resources and Services Administration (HRSA) as partial counties (census tracts and county subdivisions) or entire counties that disproportionately have a
combination of indicators that may include a low number of primary care providers per 1,000 population, high infant mortality, high poverty, and a large elderly population. Urban influence codes (UICs) are designated by the Office of Management and Budget (OMB) by population size of the metro area, and nonmetropolitan counties by size of the largest
city or town and proximity to metro and micropolitan areas (areas with a population of at least 10,000 people but fewer than 50,000). The UICs were last updated in 2013 and are updated every 10 years. Components may not sum to totals due to rounding.
Source: MedPAC analysis of HRSA MUAs, OMB UICs, and CMS enrollment data February 2020.
• In general, an MA plan’s service area consists of one or more entire counties. (MA regional
PPOs are required to cover entire regions, which consist of one or more states. In rare circumstances, MA “local” plans receive a waiver that allows them to serve only a portion of a county if the plan is able to prove that the demographic composition (e.g., income and race) of the portion of the county the plan intends to serve is not substantively different from the rest of the county.)
• We examined beneficiary access to MA plans and market share of MA plans by two geographic designations: MUAs and UICs.
(Chart continued next page)
A Data Book: Health care spending and the Medicare program, July 2020 131
Chart 9-10. The share of Medicare beneficiaries in private plans does not differ substantially in medically underserved areas compared with other areas, but is somewhat lower in rural areas, 2020 (continued)
• HRSA designates MUAs by census tract, county, or county subdivisions. HRSA designates MUAs based on a score of four combined indicators: (1) disproportionately low number of primary care providers per 1,000 people, (2) high infant mortality, (3) high poverty, and (4) a large elderly population. Part of a county may be designated as an MUA, the entire county may receive the designation, or the entire county may have no MUAs.
• The Office of Management and Budget UICs classify geographic areas as metropolitan, micropolitan, adjacent to metropolitan, and not adjacent to metropolitan; the latter three types of areas are considered rural. UICs distinguish metropolitan counties by the population size of their metro area and nonmetropolitan counties by the size of the largest city or town and proximity to metro and micropolitan areas (areas with a population of at least 10,000 people but fewer than 50,000). The UICs were last updated in 2013 and are updated every 10 years.
• Counties that have designated medically underserved areas (either partially or counties that are entirely composed of MUAs) have shares of MA enrollment similar to counties with no designated MUAs. The share of beneficiaries in MA and other private plans is the highest in counties partially designated as MUAs (41 percent). The proportion of Medicare beneficiaries in private plans located in counties that are designated entirely as MUAs (38 percent) is almost the same as counties that do not have any MUA designation (37 percent).
• Most (72 percent) of all 61.7 million Medicare beneficiaries live in metropolitan areas. The share of Medicare beneficiaries who live in metropolitan areas enrolled in MA and other private plans (43 percent) is higher than the share of rural beneficiaries enrolled in MA plans.
• Nearly all Medicare beneficiaries in rural areas reside in a micropolitan county or a county that is adjacent to a metropolitan area. Roughly one-third of Medicare beneficiaries in these areas are enrolled in MA and other private plans.
• About 1 percent of Medicare beneficiaries reside in a rural county that is not adjacent to a metropolitan area. Nearly one-quarter (23 percent) of these beneficiaries are enrolled in MA and other private plans.
132 Medicare Advantage
Chart 9-11. MA and other private plan enrollment patterns do not differ by medically underserved area designation but do vary based on urban influence designation, 2020
MA and private plan population (in millions)
As a percent of MA and
private plan population
Percent of category
HMO Local PPO
Regional PPO
Other private plans
All Medicare private plan enrollees 24.7 100% 60% 32% 5% 3% County’s medically underserved area designation Partial county 16.1 65 65 29 3 4 Entire county 4.2 17 52 35 12 2 No medically underserved areas 4.4 18 50 42 6 3 Urban influence code designation Metropolitan 19.2 78 66 28 3 3 Rural: Micropolitan 2.7 11 37 46 12 5 Rural: Adjacent to metropolitan 2.6 11 42 45 9 4 Rural: Not adjacent to metropolitan 0.2 1 28 49 13 10
Note: MA (Medicare Advantage), HMO (health maintenance organization), PPO (preferred provider organization). Beneficiaries in the Virgin Islands, Guam, American Samoa, and the Northern Mariana Islands or in foreign areas are excluded. MA
plans consist of HMOs, local PPOs, regional PPOs, private fee-for-service plans, and Medical Savings Account plans. Private plans include 0.7 million beneficiaries in non-MA private plans, which consist of Medicare–Medicaid plans, Program of All-Inclusive Care for the Elderly plans, and cost plans. Medically underserved areas (MUAs) are designated
by the Health Resources and Services Administration (HRSA) as partial counties (census tracts and county subdivisions) or entire counties that disproportionately have a combination of indicators that may include a low number of primary care providers per 1,000 population, high infant mortality, high poverty, and a large elderly population. Urban influence codes
(UICs) are designated by the Office of Management and Budget (OMB) by population size of the metro area, and nonmetropolitan counties by size of the largest city or town and proximity to metro and micropolitan areas (areas with a population of at least 10,000 people but fewer than 50,000). The UICs were last updated in 2013 and are updated every
10 years. Components may not sum to totals due to rounding. Source: MedPAC analysis of HRSA MUAs, OMB UICs, and CMS enrollment and population data February 2020.
• Local coordinated care plans (HMOs and local PPOs), which represent 92 percent of private plan enrollees, may choose which individual counties to serve. Regional PPOs (5 percent of all private plan enrollees) cover entire state-based regions.
• Enrollment by type of plan is not notably different between counties with different MUA designations. The proportion of enrollees in HMOs is similar for counties that are designated entirely as medically underserved areas (52 percent) compared with counties that do not have any medically underserved area designation (50 percent). The remainder of private plan enrollment in these areas is generally in either local or regional PPOs.
• HMOs account for the largest share of private plan enrollment in metropolitan areas (66 percent), but PPOs account for the largest share of private plan enrollment in rural areas (more than 50 percent combined between local PPOs and regional PPOs).
A Data Book: Health care spending and the Medicare program, July 2020 133
Chart 9-12. MA plans are available to nearly all beneficiaries in medically underserved and rural areas, 2020
As a percent of Medicare population
Share of Medicare beneficiaries living in counties with plans available in 2020
CCPs
Any
MA plan HMO Local PPO
HMO or local PPO
Regional PPO
Any CCP
All beneficiaries 100% 99% 95% 93% 98% 73% 99% County’s medically underserved area designation Partial county 63 99 98 93 99 69 99 Entire county 18 98 88 91 95 82 98
No medically underserved areas 19 98 94 95 97 81 98 Urban influence code designation Metropolitan 72 100 100 95 100 72 100 Rural: Micropolitan 13 97 87 90 94 71 97 Rural: Adjacent to metropolitan 14 96 84 88 92 81 96
Rural: Not adjacent to metropolitan 1 86 61 71 75 68 84
Note: MA (Medicare Advantage), CCP (coordinated care plan), HMO (health maintenance organization), PPO (preferred provider organization). These data do not include plans that have restricted enrollment or are not paid based on the MA plan bidding process (special needs plans, cost plans, employer-only plans, Program of All-Inclusive Care for the Elderly, and certain demonstration plans). Medically underserved areas (MUAs) are designated by the Health Resources and
Services Administration (HRSA) as partial counties (census tracts and county subdivisions) or entire counties that disproportionately have a combination of indicators that may include a low number of primary care providers per 1,000 population, high infant mortality, high poverty, and a large elderly population. Urban influence codes (UICs) are
designated by the Office of Management and Budget (OMB) by population size of the metro area, and nonmetropolitan counties by size of the largest city or town and proximity to metro and micropolitan areas (areas with a population of at least 10,000 people but fewer than 50,000). The UICs were last updated in 2013 and are updated every 10 years.
Components may not sum to totals due to rounding. Source: MedPAC analysis of HRSA MUAs, OMB UICs, and CMS enrollment and population data February 2020.
• We examined the availability of MA plans to all Medicare beneficiaries. Consistent with prior work, we exclude employer plans and special needs plans. Although about one-third of MA enrollees are in these excluded plans, their availability is restricted to certain populations. In addition, we do not include other private plans such as cost plans.
• MA plans are available to nearly all Medicare beneficiaries, irrespective of whether beneficiaries reside in a county with a designated medically underserved area. Among counties that are designated entirely as medically underserved areas, 98 percent of beneficiaries have access to an MA plan.
• Nearly all Medicare beneficiaries residing in metropolitan areas have access to an MA plan.
• Nearly all beneficiaries in rural counties have access to an MA plan. Between 96 percent and 97 percent of beneficiaries in micropolitan counties or those adjacent to a metropolitan area have access to an MA plan. Among the 1 percent of Medicare beneficiaries residing in a rural county that is not adjacent to a metropolitan area, 86 percent have access to an MA plan.
134 Medicare Advantage
Chart 9-13. Most Medicare beneficiaries have access to a considerable number of MA plans, but rural beneficiaries and beneficiaries in counties composed entirely of MUAs typically have fewer plans from which to choose, 2020
As a percent of Medicare population
Average plan offerings
per beneficiary
Share of Medicare beneficiaries living in
counties with an available zero-
premium plan with drug coverage
All beneficiaries 100% 27 93% County’s medically underserved area designation Partial county 63 30 93 Entire county 18 18 91 No medically underserved areas 19 25 95 Urban influence code designation Metropolitan 72 31 96 Rural: Micropolitan 13 16 82
Rural: Adjacent to metropolitan 14 16 85
Rural: Not adjacent to metropolitan 1 9 65
Note: MA (Medicare Advantage), MUA (medically underserved area). These data do not include plans that have restricted enrollment or are not paid based on the MA plan bidding process (special needs plans, cost plans, employer-only plans, and certain demonstration plans). MUAs are designated by the Health Resources and Services Administration (HRSA) as partial counties (census tracts and county subdivisions) or entire counties that disproportionately have a combination of
indicators that may include a low number of primary care providers per 1,000 population, high infant mortality, high poverty, and a large elderly population. Urban influence codes (UICs) are designated by the Office of Management and Budget (OMB) by population size of the metro area, and nonmetropolitan counties by size of the largest city or town and
proximity to metro and micropolitan areas (areas with a population of at least 10,000 people but fewer than 50,000). The UICs were last updated in 2013 and are updated every 10 years.
Source: MedPAC analysis of HRSA MUAs, OMB UICs, and CMS enrollment and population data February 2020.
• In 2020, the average beneficiary has 27 plans from which to choose in his or her county.
• On average, beneficiaries residing in counties that are designated entirely as medically underserved areas have fewer MA plans from which to choose, but still have an average of 18 plans available to them. About 91 percent of beneficiaries in these counties have a zero-premium plan with drug coverage available.
• On average, Medicare beneficiaries residing in metropolitan areas have more MA plans from which to choose (an average of 31 plan choices) compared with beneficiaries in rural areas. Nevertheless, the average beneficiary in micropolitan counties or those adjacent to a metropolitan area can choose among an average of 16 plans. Beneficiaries residing in rural counties that are not adjacent to a metropolitan area (1 percent of all beneficiaries) have 9 plans from which to choose, on average.
• At least one zero-premium plan with drug coverage is available to most beneficiaries (93 percent). Availability of these plans in rural areas is somewhat less prevalent than in metropolitan areas. In metropolitan areas, 96 percent of beneficiaries have access to a zero-premium plan. In comparison, over 80 percent of beneficiaries in micropolitan counties or those adjacent to a metropolitan area have access to a zero-premium plan. In rural counties that are not adjacent to a metropolitan area, 65 percent of beneficiaries have an available zero-premium plan.
A Data Book: Health care spending and the Medicare program, July 2020 135
Chart 9-14. Twenty most common condition categories among MA beneficiaries, as defined in the CMS–HCC model, 2018
Percent of Percent of beneficiaries beneficiaries with listed condition Conditions (defined by HCC) with listed condition and no others Diabetes with chronic complications 19.7% 3.6%
Vascular disease 18.7 2.2
COPD 14.0 1.7
CHF 11.6 0.5
Specified heart arrhythmias 11.4 1.3
Major depressive, bipolar, and paranoid disorders 11.3 1.9
Diabetes without complications 8.5 3.1
Morbid obesity 8.5 1.0
Rheumatoid arthritis and inflammatory connective tissue disease 6.5 1.1
Breast, prostate, colorectal, and other cancers and tumors 5.1 1.3
Coagulation defects and other specified hematological disorders 4.8 0.4
Angina pectoris 4.0 0.3
Drug/alcohol dependence 3.6 0.3
Other significant endocrine and metabolic disorders 3.5 0.3
Acute renal failure 3.4 0.1
Cardio-respiratory failure and shock 2.5 0.0
Seizure disorders and convulsions 2.5 0.3
Ischemic or unspecified stroke 2.2 0.1
Septicemia, sepsis, systemic inflammatory response
syndrome/shock 1.8 0.0
Hemiplegia/Hemiparesis 1.6 0.1
Note: MA (Medicare Advantage), CMS–HCC (CMS–hierarchical condition category), COPD (chronic obstructive pulmonary
disease), CHF (congestive heart failure).
Source: MedPAC analysis of Medicare data files from Acumen LLC.
• CMS uses the CMS–HCC model to risk adjust capitated payments to MA plans so that payments better reflect the clinical needs of MA enrollees given the number and severity of their clinical conditions. The CMS–HCC model uses beneficiaries’ conditions, which are collected into HCCs, to adjust the capitated payments.
• Diabetes with chronic complications is the most common HCC, and over 28 percent of MA enrollees are in at least one of the two diabetes HCCs.
136 Medicare Advantage
Chart 9-15. Medicare private plan enrollment patterns, by age and Medicare–Medicaid dual-eligible status, December 2018
As percent of
Medicare population Percent of
category in FFS
Percent of category in
private plans
All beneficiaries 100% 65% 35% Aged (65 or older) 85 64 36 Under 65 15 68 32 Non–dual eligible 82 66 34 Aged (65 or older) 74 65 35 Under 65 8 69 31 Dual eligible 18 60 40 Aged (65 or older) 11 56 44
Under 65 7 67 33
Dual-eligible beneficiaries by category (all ages)
Full dual eligibility 13 65 35 Beneficiaries with partial dual eligibility QMB only 3 53 47 SLMB only 2 48 52 QI 1 46 54
Note: FFS (fee-for-service), QMB (qualified Medicare beneficiary), SLMB (specified low-income beneficiary), QI (qualified individual). The Medicare population includes beneficiaries who have either Part A coverage or Part B coverage. Dual-eligible beneficiaries are eligible for Medicare and Medicaid. See accompanying text for an explanation of the categories
of dual-eligible beneficiaries. “Plans” include Medicare Advantage plans as well as cost-reimbursed plans. Data exclude Puerto Rico because of the inability to determine specific dual-eligible categories. As of December 2018, Puerto Rico had 579,000 Medicare beneficiaries enrolled in private plans. Dual-eligible special needs plans in Puerto Rico had 281,000
enrollees in December 2018. Figures may not sum to totals due to rounding. Source: MedPAC analysis of 2018 denominator and common Medicare environment files and CMS monthly Medicare Advantage
reports.
• Medicare plan enrollment among the dually eligible continues to increase. In 2018, 40 percent of dual-eligible beneficiaries were in Medicare private plans, up from 36 percent in 2017.
• A substantial share of dual-eligible beneficiaries (40 percent (not shown in table)) are under the age of 65 and entitled to Medicare on the basis of disability or end-stage renal disease. Regardless of dual-eligibility status, beneficiaries under age 65 are less likely than aged beneficiaries to enroll in Medicare private plans (32 percent vs. 36 percent, respectively).
• Dual-eligible beneficiaries who have full dual eligibility—that is, those who have coverage for their Medicare out-of-pocket costs (premiums and cost sharing) as well as coverage for services such as long-term care services and supports—are less likely to enroll in private Medicare plans than beneficiaries with “partial” dual eligibility. Full dual-eligibility categories consist of beneficiaries with coverage through state Medicaid programs as well as certain QMBs and SLMBs who also have Medicaid coverage for services. The latter two categories are referred to as QMB-Plus and SLMB-Plus beneficiaries. Beneficiaries with partial dual eligibility have coverage for Medicare premiums (through the QI or SLMB program) or premiums and Medicare cost sharing, in the case of the QMB program. SLMB-only and QI beneficiaries have higher rates of plan enrollment (52 percent and 54 percent, respectively) than any other category shown in this chart, and the rates are higher than the average rate (35 percent) across all Medicare beneficiaries.
10
00
A Data Book: Health care spending and the Medicare program, July 2020 139
Chart 10-1. Medicare spending for Part B drugs furnished by physicians, hospital outpatient departments, and suppliers, 2005–2018
Note: HOPD (hospital outpatient department). Data include Part B–covered drugs furnished by several provider types, including
physicians, suppliers, and hospital outpatient departments, and exclude those furnished by critical access hospitals,
Maryland hospitals, and dialysis facilities. “Medicare spending” includes program payments and beneficiary cost sharing. Data reflect all Part B drugs whether they were paid based on the average sales price or another payment formula. Data exclude blood and blood products (other than clotting factor). Components may not sum to totals due to rounding.
Source: MedPAC and Acumen LLC analysis of Medicare claims data.
• The Medicare program and beneficiaries spent about $35 billion on Part B drugs furnished by physicians, hospital outpatient departments (HOPDs), and suppliers in 2018, an increase of about 9 percent from 2017.
• Since 2005, Medicare pays for most Part B drugs at a rate of the average sales price plus 6 percent (ASP + 6 percent). Between 2005 and 2018, total spending grew at an average annual rate of 7.7 percent. Spending growth was slower from 2005 to 2009 (about 3.7 percent per year on average) and more rapid from 2009 to 2018 (about 9.5 percent per year on average).
• Eligible hospitals that participate in the 340B drug discount program receive substantial discounts on outpatient drugs, including those covered by Medicare Part B. Beginning 2018, Medicare reduced the payment rate for certain Part B drugs furnished by 340B hospitals to ASP – 22.5 percent.
• Spending on Part B–covered drugs in outpatient hospitals was $13.1 billion in 2018. If the 340B payment reduction had not occurred, we estimate that Part B drug spending in outpatient hospitals would have been $1.8 billion higher (that is, approximately $15 billion) in 2018.
(Chart continued next page)
13.4 14.0 14.9 14.5
15.4 16.6
18.6 20.3
21.6 22.8
25.8
29.2
32.1
35.0
9.1 9.6 9.9 9.8 10.2 10.711.6 12.2 12.6 13.2
15.016.6
18.019.9
2.8 2.9 3.2 3.1 3.6 4.2 5.2
6.1 6.9 7.5 8.7
10.412.3 13.1
1.4 1.5 1.8 1.6 1.7 1.7 1.8 2.0 2.1 2.2 2.1 2.1 1.8 1.90
3
6
9
12
15
18
21
24
27
30
33
36
39
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Med
icare
sp
en
din
g (d
oll
ars
in
bil
lio
ns) Total
Physicians
HOPDs
Suppliers
140 Prescription drugs
Chart 10-1. Medicare spending for Part B drugs furnished by physicians, hospital outpatient departments, and suppliers, 2005–2018 (continued)
• Of total 2018 Part B drug spending, physicians accounted for 57 percent ($19.9 billion), HOPDs accounted for 38 percent ($13.1 billion), and suppliers accounted for 5 percent ($1.9 billion).
• Overall, from 2009 and 2018, Part B drug spending has grown more rapidly for HOPDs than for physicians and suppliers—at average annual rates of about 15 percent, 8 percent, and 2 percent, respectively. However, in the most recent one-year period from 2017 to 2018, Part B drug spending grew more rapidly in physician offices (11 percent) than HOPDs (7 percent) because Medicare reduced the payment rates for some Part B drugs furnished by 340B hospitals beginning 2018. If 340B hospitals had been paid ASP + 6 percent instead of ASP – 22.5 percent, we estimate that Part B drug spending in HOPDs would have grown 22 percent between 2017 and 2018 instead of 7 percent.
• Between 2017 and 2018, Medicare payments for supplier-furnished drugs (such as inhalation drugs, home infusion drugs, and three types of oral drugs) increased 4 percent. In the prior year—from 2016 to 2017—spending on supplier-furnished drugs decreased as a result of a change in the payment formula for Part B–covered home infusion drugs (from 95 percent of the average wholesale price to ASP + 6 percent) and patent expirations and generic entry for certain products.
• Not included in these data are critical access hospitals and Maryland hospitals, which are not paid under the ASP system, and end-stage renal disease facilities, which are paid for most Part B drugs through the dialysis bundled payment rate. Medicare and beneficiaries spent approximately $930 million in critical access hospitals and $380 million in Maryland hospitals for Part B drugs in 2018. In addition, in 2018, Medicare spent $1.2 billion for calcimimetics in dialysis facilities through a transitional drug add-on payment adjustment to the bundled dialysis payment rate.
A Data Book: Health care spending and the Medicare program, July 2020 141
Chart 10-2. Change in Medicare payments and utilization for separately payable Part B drugs, 2009–2018
2009
2018
Average annual growth
2009–2018
Total payments: Separately payable Part B drugs (in billions) $11.9* $33.1* 12.0%
Total payments: All Part B drugs excluding vaccines (in billions) $11.7 $31.8 11.7
Number of beneficiaries using a Part B drug (in millions) 2.7 3.9 4.3 Average total payments per beneficiary who used a Part B drug $4,402 $8,165 7.1
Average number of Part B drugs per beneficiary 1.39 1.35 –0.4 Average annual payment per Part B drug per beneficiary $3,158 $6,047 7.5
Total payments: All Part B vaccines (in billions) $0.2 $1.3 21.9
Number of beneficiaries using a Part B vaccine (in millions) 13.4 16.8 2.5 Average total payments per beneficiary who used a Part B vaccine $16 $77 18.9 Average number of Part B vaccines per beneficiary 1.08 1.20 1.2 Average annual payment per Part B vaccine per beneficiary $15 $64 17.5
Note: This analysis includes Part B drugs paid based on the average sales price as well as the small group of Part B drugs that
are paid based on the average wholesale price or reasonable cost or that are contractor priced. “Vaccines” refers to the
three Part B–covered preventive vaccines: influenza, pneumococcal, and hepatitis B. Data include Part B drugs furnished by physicians, hospitals paid under the outpatient prospective payment system, and suppliers and exclude data for critical access hospitals, Maryland hospitals, and dialysis facilities. Yearly figures presented in the table are rounded; the average
annual growth rate was calculated using unrounded data. *For purposes of this analysis, spending on separately payable Part B drugs excludes any drug that was bundled in 2009 or
2018 (i.e., drugs that were packaged under the outpatient prospective payment system in 2009 or 2018 were excluded from both years of the analysis, regardless of the setting where the drug was administered), drugs billed under not-otherwise-classified billing codes, and blood and blood products (other than clotting factor). Without those exclusions, Part B drug
spending was $15.4 billion in 2009 and $35.0 billion in 2018, as shown in Chart 10-1. Source: MedPAC analysis of Medicare claims data for physicians, hospital outpatient departments, and suppliers.
• Total payments by the Medicare program and beneficiaries for separately payable Part B drugs increased 12.0 percent per year, on average, between 2009 and 2018.
• Medicare spending on separately payable Part B drugs excluding Part B–covered preventive vaccines grew at a similar rate (11.7 percent per year) between 2009 and 2018.
• The largest factor contributing to the growth in separately payable Part B drug spending (excluding vaccines) was the change in the price Medicare paid for drugs. Between 2009 and 2018, the average annual payment per drug increased on average by 7.5 percent per year, which reflects increases in the prices of existing drugs and changes in the mix of drugs, including the adoption of new, higher priced drugs. Growth in the average payment per drug would have been even higher if not for the 2018 reduction in Medicare’s payment rate for certain Part B drugs provided by 340B hospitals.
(Chart continued next page)
142 Prescription drugs
Chart 10-2. Change in Medicare payments and utilization for separately payable Part B drugs, 2009–2018 (continued)
• Growth in the number of beneficiaries using nonvaccine Part B drugs (about 4.3 percent per year on average) also contributed to increased spending. The number of Part B drugs received per user declined from about 1.39 in 2009 to 1.35 in 2018, which modestly offset spending growth.
• Medicare covers three preventive vaccines: influenza, pneumococcal, and—for beneficiaries at high or medium risk—hepatitis B. Spending on the three preventive vaccines furnished by physicians, hospital outpatient departments, and pharmacy suppliers was $690 million for influenza, $617 million for pneumococcal, and $7 million for hepatitis B. (Not included in these data are vaccines furnished in other settings such as ESRD facilities. With other settings included, 2018 vaccine spending was $706 million on influenza, $627 million on pneumococcal, and $38 million on hepatitis B vaccines.)
• Although vaccines are a relatively small share of overall spending on separately payable Part B drugs, vaccine spending grew rapidly, at an average rate of about 22 percent per year, between 2009 and 2018.
• As with other separately paid Part B drugs, the largest driver of increased vaccine spending was price growth, as the average payment per vaccine grew at an average rate of 17.5 percent per year between 2009 and 2018. Substantial price growth occurred for both pneumococcal and flu vaccines between 2009 and 2018, with the average payment per vaccine increasing from $36 to $150 for pneumococcal and from $12 to $42 for flu vaccines over this period (data not shown). In addition, use of the pneumococcal vaccine Prevnar-13 increased following a 2014 recommendation by the Centers for Disease Control and Prevention advisory committee for a one-time vaccination of all people age 65 and older. Because the price of Prevnar-13 is higher than other Part B–covered vaccines, its increased use has contributed to growth in the average payment per vaccine.
A Data Book: Health care spending and the Medicare program, July 2020 143
Chart 10-3. Top 10 Part B drugs paid based on ASP, by type of provider, 2017 and 2018
Dollars (in millions)
Total Physician and supplier HOPD Part B drug spending Part B drug spending Part B drug spending
2017 2018 2017 2018 2017 2018
Eylea $2,469 $2,577 $2,312 $2,435 $158 $142
Keytruda 1,037 1,813 394 764 643 1,049
Opdivo 1,474 1,718 695 827 778 891
Rituxan 1,758 1,703 857 867 900 836
Prolia/Xgeva 1,243 1,420 763 909 481 511
Neulasta 1,405 1,373 653 640 751 733
Lucentis 1,039 1,217 1,006 1,186 32 30
Remicade 1,347 1,154 821 745 526 408
Avastin 1,071 1,014 524 503 547 511
Herceptin 786 823 354 386 432 438
Total spending, top 10 drugs $13,627 $14,812 $8,379 $9,263 $5,249 $5,549
Total spending, all Part B drugs $32,083 $34,955 $19,801 $21,832 $12,282 $13,123
Note: ASP (average sales price), HOPD (hospital outpatient department). The 10 drugs shown in the chart reflect the Part B drug billing codes paid under the ASP methodology with the highest Medicare expenditures in 2018. Data for 2017 are shown for comparison. Data include Part B–covered drugs furnished by several provider types, including physicians,
suppliers, and hospital outpatient departments, but exclude those furnished by critical access hospitals, Maryland hospitals, and dialysis facilities. “Drug spending” includes Medicare program payments and beneficiary cost sharing. “Total spending, all Part B drugs” reflects all products, whether paid based on ASP or another method. Data exclude
blood and blood products (other than clotting factor). Components may not sum to totals due to rounding. Source: MedPAC and Acumen LLC analysis of Medicare claims data.
• Part B drugs are billed under more than 700 billing codes, but spending is concentrated. Medicare spending (including cost sharing) on the top 10 drugs paid under the ASP system totaled about $14.8 billion in 2018, about 42 percent of all Part B drug spending that year.
• As of 2018, all of the top 10 Part B drugs are biologics. Many of these products are used to treat cancer or its side effects (Keytruda, Opdivo, Rituxan, Prolia/Xgeva, Neulasta, Avastin, Herceptin). Drugs used to treat age-related macular degeneration (Eylea, Lucentis, Avastin) and rheumatoid arthritis (Remicade and Rituxan) are also in the top 10.
• Medicare spending on immune globulin (for which there are several products billed through separate billing codes) amounted to more than $1.5 billion in 2018 (data not shown).
144 Prescription drugs
Chart 10-4. Growth in ASP for the 20 highest expenditure Part B drugs, 2005–2020
Part B drug
Total Medicare payments in 2018
(in billions)
Average annual ASP growth Earliest year of
ASP data if not 2005
2005– 2015
2015– 2019
2019– 2020
2005– 2020
Eylea 2.6 0.0%* –0.4% –1.9% –0.5%* 2013
Keytruda 1.8 N/A 2.3* 2.6 2.4* 2016
Opdivo 1.7 N/A 2.7* 2.4 2.6* 2016
Rituxan 1.7 5.1 7.1 –1.0 5.2
Prolia/Xgeva 1.4 0.6* 6.1 3.6 3.7* 2012
Neulasta 1.4 4.4 7.6 –9.2 4.3
Lucentis 1.2 –0.4* –2.1 –4.6 –1.3* 2008
Remicade 1.2 3.4 0.8 –25.2 0.5
Avastin 1.0 1.8 4.6 –0.7 2.3
Herceptin 0.8 4.8 6.1 –1.1 4.7
Orencia 0.8 7.4* 11.7 6.3 8.6* 2007
Darzalex 0.6 N/A 5.6* 3.3 4.8* 2017
Ocrevus 0.5 N/A 0.5* 0.1 0.3* 2018
Alimta 0.5 4.1 2.9 3.9 3.8
Velcade 0.4 5.1 –0.5 –1.7 3.1
Sandostatin LAR 0.4 5.3 8.4 4.1 6.1
Xolair 0.4 6.1 7.6 0.8 6.1
Soliris 0.4 2.5* 2.3 0.0 2.3* 2008
Gammagard 0.4 3.1* 1.9 –5.5 2.0* 2008
Gamunex-C / Gammaked 0.4 2.9* –0.1 0.6 1.7* 2008
Consumer price index for urban consumers 2.1 1.9 2.5 2.0 Note: ASP (average sales price), N/A (not applicable). Growth rates for ASP are calculated from first quarter to first quarter of each
year. “Medicare payments” includes Medicare program payments and beneficiary cost sharing for these drugs furnished by
physicians, suppliers, and hospital outpatient departments, but excludes those furnished by critical access hospitals, Maryland hospitals, and dialysis facilities. Vaccines for which Medicare pays 95 percent of the average wholesale price are also excluded from this table. See Chart 10-2 and associated bullets for information on vaccine price growth.
*Indicates that ASP payment rates for a specific product were not available for the full period listed, and the average annual growth rate was calculated based on the earliest year that a first-quarter payment rate was available.
Source: MedPAC analysis of CMS ASP pricing files and consumer price index for all urban consumers data from the Bureau of
Labor Statistics and MedPAC and Acumen LLC analysis of Medicare claims data.
• Over the period from 2005 to 2020, 18 out of 20 of the top Part B drugs have experienced price increases, with 14 of these products’ ASPs increasing faster than the consumer price index for urban consumers.
(Chart continued next page)
A Data Book: Health care spending and the Medicare program, July 2020 145
Chart 10-4. Growth in ASP for the 20 highest expenditure Part B drugs, 2005–2020 (continued)
• In the most recent year, price changes have been mixed. Among the top 20 Part B drugs,
ASP increased for 10 products, decreased for 9 products, and was unchanged for 1 product between the first quarters of 2019 and 2020.
• Biosimilar competition may account for the decreases in ASP between 2019 and 2020 for some originator biologics, as Rituxan, Neulasta, Remicade, Avastin, and Herceptin all faced biosimilar entry during 2019 or earlier. For these five products, the recent price declines begin to reverse a long period of rising prices, with average price growth over the last 15 years ranging from 0.5 percent per year for Remicade to 5.2 percent per year for Rituxan.
146 Prescription drugs
Chart 10-5. Trends in Medicare Part B payment rates for originator biologics and their biosimilar products
2016 Q1
2017 Q1
2018 Q1
2018 Q3
2019 Q1
2019 Q3
2020 Q1
Neupogen and biosimilars
Neupogen’s payment rate $1.01 $1.00 $1.00 $1.02 $1.00 $0.98 $0.96
Percent biosimilars’ payment rates are below Neupogen’s payment rate 4–24% 22–29% 31–39% 36–42% 25–42% 30–41% 34–45%
Biosimilar market share 25% 51% 63% 67% 70% 73% N/Aa
Remicade and biosimilars
Remicade’s payment rate $79.91 $82.22 $85.81 $83.90 $76.65 $64.87 $57.35
Percent biosimilars’ payment rates are below Remicade’s payment rate N/Ab –22% 12% 17–23% 19–25% 16–21% 10–17%
Biosimilar market share N/Ab 0% 6% 9% 11% 14% N/Aa
Neulasta and biosimilars
Neulasta’s payment rate $3,828 $4,117 $4,442 $4,721 $4,682 $4,528 $4,252
Percent biosimilars’ payment rates are below Neulasta’s payment rate N/Ab N/Ab N/Ab N/Ac 8% 5–12% 5–12%
Biosimilar market share N/Ab N/Ab N/Ab 1% 8% 20% N/Aa
Procrit/Epogen and biosimilars
Procrit/Epogen payment rate $12.33 $12.32 $12.13 $13.09 $11.95 $11.46 $10.56
Percent biosimilar’s payment rate is below Procrit/ Epogen’s payment rate N/Ab N/Ab N/Ab N/Ab 5% 11% 12%
Biosimilar market share N/Ab N/Ab N/Ab N/Ab 6% 24% N/Aa
Note: Q1 (first quarter), Q3 (third quarter), N/A (not available). An originator biologic is a drug product derived from a living
organism. A biosimilar product is a follow-on product that is approved by the Food and Drug Administration based on the
product being highly similar to the originator biologic. The biosimilars included in the analysis are Zarxio, Nivestym, and Granix for originator biologic Neupogen; Inflectra and Renflexis for originator biologic Remicade; Fulphila and Udenyca for originator Neulasta; and Retacrit for originator Procrit//Epogen. Although Granix is not a biosimilar in the U.S. (because it
was approved under the standard Food and Drug Administration approval process for new biologics), we include it here because it was approved as a biosimilar to Neupogen in Europe and it functions as a competitor to Neupogen in the U.S. market. For Remicade, the biosimilar’s payment rate was 22 percent above the originator’s payment rate in first quarter
2017. aClaims data on utilization are not yet available for this quarter. bBiosimilar product was not yet approved and/or launched.
cA published payment rate was unavailable for this product for this quarter. Source: MedPAC analysis of payment rates from CMS ASP pricing files. MedPAC and Acumen LLC analysis of Medicare claims data.
(Chart continued next page)
A Data Book: Health care spending and the Medicare program, July 2020 147
Chart 10-5. Trends in Medicare Part B payment rates for originator biologics and their biosimilar products (continued)
• An originator biologic is a product derived from a living organism. A biosimilar product is a follow-on product approved by the Food and Drug Administration based on the product being highly similar to the originator biologic.
• Under Part B, Medicare pays for an originator biologic at 106 percent of its own average sales price (ASP). For biosimilars, Medicare pays 100 percent of the biosimilar’s ASP + 6 percent of the originator product’s ASP.
• Medicare payment rates for biosimilars are lower than those of the corresponding originator biologics because biosimilars have lower prices (as measured by ASP) than originator biologics. The extent to which originator biologics have lowered their prices in response to biosimilar entry and the extent to which market share has shifted to biosimilars varies by product.
• For Neupogen, the originator biologic that has faced biosimilar competition for the longest period (since late 2015), the payment rates for biosimilar products are substantially less (34 percent to 45 percent less as of the first quarter of 2020) than for the originator. The originator Neupogen has reduced its price only modestly (5 percent) between 2016 and 2020. Biosimilars account for the majority of utilization: 73 percent market share as of the third quarter of 2019.
• For Remicade, the payment rates for biosimilar products are about 10 percent to 17 percent below the originator’s payment rate as of the first quarter of 2020. After biosimilar entry in late 2016, the originator Remicade’s price initially increased (4 percent between the first quarters of 2017 and 2018). Subsequently Remicade’s price decreased substantially, falling 33 percent between the first quarters of 2018 and 2020. Remicade has continued to retain most of the market share, with biosimilars accounting for only 14 percent of utilization as of the third quarter of 2019.
• For Neulasta, which has faced biosimilar competition since mid-2018, biosimilars’ payment rates are 5 percent to 12 percent below the originator’s payment rate as of the first quarter of 2020. The originator Neulasta’s price has decreased 10 percent between the third quarter of 2018 and the first quarter of 2020. Biosimilar utilization is growing, reaching a market share of 20 percent as of the third quarter of 2019.
• For Procrit/Epogen, which have faced biosimilar competition since late 2018, the payment rate of their biosimilar is 12 percent below the originator’s payment rate as of the first quarter of 2020. The originator’s payment rate has decreased 19 percent between the third quarter of 2018 and the first quarter of 2020. The biosimilar accounts for almost one-fourth of utilization as of the third quarter of 2019.
• Not shown in the chart, three additional originator products—Avastin, Herceptin, and Rituxan—faced biosimilar entry during the second half of 2019. The biosimilars’ payment rates are 13 percent (Avastin), 14 percent (Herceptin), and 8 percent (Rituxan) below the originator’s payment rate as of the first quarter of 2020 (data not shown). The originator products’ ASPs have declined by roughly 1 percent between the first quarter of 2019 and 2020 (as shown in Chart 10-4).
148 Prescription drugs
Chart 10-6. Price indexes for Medicare Part B drugs, 2005–2018
Note: Q1 (first quarter), Q4 (fourth quarter). The Part B price indexes reflect growth in the average sales price of Part B–covered
drugs over time, measured for individual drugs at the level of the Healthcare Common Procedure Coding System billing code. These measures of price growth reflect growth in the price of individual products but do not reflect changes in price due
to the introduction of new products or changes in the mix of products used. The Part B price index for biologics in this chart and in Chart 10-27 are different due to the different periods of analysis.
Source: Acumen LLC analysis for MedPAC. • The Part B price indexes reflect growth in the average sales price (ASP) at the individual
product level and do not reflect changes in price that occur as a result of changes in the mix of drugs used or the introduction of new, higher priced drugs.
• Measured by the change in the ASP of individual Part B–covered drugs, the prices of Part B–covered drugs rose by an average of about 18 percent cumulatively between 2005 and 2018 (an index of 1.18).
• Underlying this overall trend in the price index are different patterns by type of product. The price index for Part B–covered biologics increased by 46 percent between 2005 and 2018 (an index of 1.46). In contrast, the price index for nonbiologics declined by 19 percent (an index of 0.81) over this period. The nonbiologic group includes single-source drugs and drugs with generic competition. The downward price trend for nonbiologics in part reflects patent expiration and generic entry for some of these products. It also reflects the design of the ASP payment system, which spurs price competition among generics and their associated brand-name products by assigning these products to a single billing code and paying them the same average rate.
1.18
1.46
0.81
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
2005Q1
2006Q1
2007Q1
2008Q1
2009Q1
2010Q1
2011Q1
2012Q1
2013Q1
2014Q1
2015Q1
2016Q1
2017Q1
2018Q1
2018Q4
Ch
ain
-weig
hte
d F
ish
er
pri
ce i
nd
ex
Biologics
Nonbiologics
All Part B drugs
A Data Book: Health care spending and the Medicare program, July 2020 149
Chart 10-7. In 2018, 88 percent of Medicare beneficiaries were enrolled in Part D plans or had other sources of creditable drug coverage
Note: LIS (low-income [drug] subsidy), PDP (prescription drug plan), MA–PD (Medicare Advantage–Prescription Drug [plan]), RDS (retiree drug subsidy). “Creditable coverage” means the value of drug benefits is equal to or greater than that of the basic Part D benefit. Components may not total to 100 due to rounding.
Source: MedPAC analysis of the Medicare denominator file 2018.
• In 2018, 88 percent of Medicare beneficiaries were enrolled in Part D plans, got prescription drug coverage through employer-sponsored plans that received Medicare’s RDS, or had other sources of drug coverage that were equal to or greater than the average value of Part D’s defined standard benefit (called “creditable coverage”). Twelve percent of Medicare beneficiaries had no drug coverage or coverage that was less generous.
• That same year, nearly three-quarters of Medicare beneficiaries received prescription drug benefits through Part D plans: 43 percent in stand-alone PDPs and 31 percent in MA–PDs.
• Nearly 22 percent of Medicare beneficiaries received Part D’s LIS in 2018. Of all LIS beneficiaries, about three-fifths of them (13 percent of all Medicare beneficiaries) were enrolled in stand-alone PDPs, and the remaining beneficiaries (8 percent) were in MA–PDs.
• Other enrollees in stand-alone PDPs accounted for 30 percent of all Medicare beneficiaries. Another 23 percent of Medicare beneficiaries were enrolled in MA–PDs and did not receive low-income subsidies.
(Chart continued next page)
Non-LIS enrollees
in PDPs30%
LIS enrollees in
PDPs13%
Non-LIS enrollees
in MA–PDs23%
LIS enrollees in
MA–PDs8%
Primary coverage
through employers that received RDS
2%
Other sources of
creditable coverage11%
No creditable
coverage12%
PDPs43%
MA‒PDs 31%
150 Prescription drugs
Chart 10-7. In 2018, 88 percent of Medicare beneficiaries were enrolled in Part D plans or had other sources of creditable drug coverage (continued)
• Employer and union health plans continue to be important sources of drug coverage for Medicare beneficiaries. In 2018, 12 percent of Medicare beneficiaries were in Part D plans (including PDPs and MA–PDs) set up by employers or unions for their retirees (data not shown). Under these employer group waiver plans (EGWPs), Medicare is the primary payer for basic drug benefits, and typically the employer offers wrap-around coverage. Separately, 2 percent of Medicare beneficiaries were in plans offered by employers that received Medicare’s RDS. (If an employer remains the primary payer of creditable drug coverage for its retirees, Medicare provides the employer with a tax-free subsidy for 28 percent of each eligible individual’s drug costs that fall within a specified range of spending.) Additionally, in 2018, 11 percent of Medicare beneficiaries had creditable drug coverage from sources other than Part D, much (but not all) of which was related to past employment, for example, through the Federal Employees Health Benefits Program, TRICARE, and employers that did not sponsor an EGWP or receive the RDS.
A Data Book: Health care spending and the Medicare program, July 2020 151
Chart 10-8. Changes in parameters of the Part D defined standard benefit over time
Cumulative change 2006 2018 2019 2020 2006–2020 Deductible $250.00 $405.00 $415.00 $435.00 74%
Initial coverage limit 2,250.00 3,750.00 3,820.00 4,020.00 79%
Annual out-of-pocket threshold 3,600.00 5,000.00 5,100.00 6,350.00 76%
Total covered drug spending at annual out-of-pocket threshold 5,100.00 8,417.60 8,139.54 9,719.38 91%
Cost sharing above the annual out-of-pocket threshold is the greater of 5% coinsurance or these amounts:
Copay for generic/preferred multisource drugs 2.00 3.35 3.40 3.60 80%
Copay for other prescription drugs 5.00 8.35 8.50 8.95 79% Note: Under Part D’s defined standard benefit, the enrollee pays the deductible and then 25 percent of covered drug spending
(75 percent is paid by the plan) until total covered drug spending reaches the initial coverage limit (ICL). Before 2011, enrollees exceeding the ICL were responsible for 100 percent of covered drug spending up to the annual out-of-pocket (OOP) threshold. Beginning in 2011, enrollees pay reduced cost sharing in the coverage gap. For 2011 and later years,
the amount of total covered drug spending at the annual OOP threshold depended on the mix of brand-name and generic drugs filled during the coverage gap. The amounts shown are for individuals not receiving Part D’s low-income subsidy who have no source of supplemental coverage. Cost sharing paid by most sources of supplemental coverage does not
count toward this threshold. The amount for 2019 is lower than that for 2018 because of a change in law that causes 95 percent of an enrollee’s spending for brand-name drugs in Part D’s coverage-gap phase to count toward the OOP threshold, compared with 85 percent in 2018. Above the OOP limit, the enrollee pays 5 percent coinsurance or the
respective copay shown above, whichever is greater. Source: CMS Office of the Actuary.
• The Medicare Prescription Drug, Improvement, and Modernization Act of 2003 specified a defined standard benefit structure for Part D. In 2020, the standard benefit has a $435 deductible, 25 percent coinsurance on covered drugs until the enrollee reaches $4,020 in total covered drug spending, and then a coverage gap until OOP spending reaches the annual threshold. (The total dollar amount of drug spending at which a beneficiary reaches the OOP threshold varies from person to person, depending on the mix of brand-name and generic prescriptions filled. CMS estimates that in 2020, a person who does not receive Part D’s low-income subsidy and has no supplemental coverage would, on average, reach the threshold at $9,719.38 in total drug spending.) Before 2011, enrollees were responsible for paying the full discounted price of drugs filled during the coverage gap. Because of changes made by the Affordable Care Act of 2010, enrollees pay reduced cost sharing for drugs filled in the coverage gap. In 2020, the cost sharing for drugs filled during the gap phase is about 25 percent for brand-name drugs and generics. Enrollees with drug spending that exceeds the annual threshold pay the greater of $3.60 to $8.95 or 5 percent coinsurance per prescription.
(Chart continued next page)
152 Prescription drugs
Chart 10-8. Changes in parameters of the Part D defined standard benefit over time (continued)
• Most parameters of this defined standard benefit structure have changed over time at the same rate as the annual change in average total drug expenses of Medicare beneficiaries enrolled in Part D, with cumulative changes of 74 percent to 79 percent between 2006 and 2020. The out-of-pocket threshold for 2020 is much higher than that for 2019 because the 2019 amount was restrained by a provision in law that limited increases between 2014 and 2019. In 2020, the out-of-pocket threshold reverted to what it otherwise would have been had CMS increased it by the same factor as other benefit parameters—that is, annual growth in Part D spending per enrollee. The effects of this increase on beneficiaries are somewhat muted by the fact that manufacturers provide a 70 percent discount on brand-name drugs in the coverage-gap phase, which counts as beneficiary spending toward the threshold.
• Within certain limits, sponsoring organizations may offer Part D plans that have the same actuarial value as the defined standard benefit but a different benefit structure, and most sponsoring organizations do offer such plans. For example, a plan may use tiered copayments rather than 25 percent coinsurance or have no deductible but use cost-sharing requirements that are equivalent to a rate higher than 25 percent. Defined standard benefit plans and plans that are actuarially equivalent to the defined standard benefit are both known as “basic benefits.”
• Once a sponsoring organization offers one plan with basic benefits within a prescription drug plan region, it may also offer a plan with enhanced benefits—basic and supplemental coverage combined.
• Under the Bipartisan Budget Act of 2018, manufacturers of brand-name drugs must provide a 70 percent discount in the coverage gap, enrollees pay 25 percent cost sharing, and plan sponsors are responsible for covering only 5 percent of the cost of brand-name drugs.
A Data Book: Health care spending and the Medicare program, July 2020 153
Chart 10-9. Characteristics of stand-alone Medicare PDPs 2019 2020
Enrollees as of Enrollees as of Plans February 2019 Plans February 2020
Number Number Number Percent (in millions) Percent Number Percent (in millions) Percent Total 901 100% 20.8 100% 948 100% 20.5 100%
Type of organization National 746 83 19.4 93 716 76 18.8 92 Other 155 17 1.4 7 232 24 1.7 8
Type of benefit Defined standard 0 0 0.0 0 0 0 0.0 0 Actuarially equivalent 348 39 12.1 58 382 40 11.3 55 Enhanced 553 61 8.7 42 566 60 9.2 45
Type of deductible Zero 263 29 8.1 39 133 14 3.0 15 Reduced 170 19 3.3 16 161 17 5.0 25 Defined standard* 468 52 9.4 45 654 69 12.4 61
Some formulary tiers not subject to a deductible Some 414 46 8.2 39 504 53 11.5 56 None 487 54 12.6 61 444 47 9.0 44
Note: PDP (prescription drug plan). The PDPs and enrollment described here exclude employer-only plans and plans offered in
U.S. territories. “National” data reflect the total number of plans for organizations with at least 1 PDP in each of the 34
PDP regions. “Actuarially equivalent” includes both actuarially equivalent standard and basic alternative benefits. “Enhanced” refers to plans with basic plus supplemental coverage. Components may not sum to totals due to rounding.
*The defined standard benefit’s deductible was $415 in 2019 and is $435 in 2020.
Source: MedPAC analysis of CMS landscape, premium, and enrollment data.
• Plan sponsors are offering 948 stand-alone PDPs in 2020 compared with 901 in 2019—an increase of more than 5 percent. Total enrollment in PDPs declined by 1.5 percent to 20.5 million beneficiaries in 2020 from 20.8 million in 2019.
• In 2020, 76 percent of all PDPs are offered by sponsoring organizations that have at least 1 PDP in each of the 34 PDP regions (shown as “national” organizations in the table). Plans offered by those national sponsors account for 92 percent of all PDP enrollment.
• For 2020, 60 percent of PDP offerings include enhanced benefits (basic plus supplemental coverage), nearly the same percentage as in 2019. In 2020, the share of PDPs with actuarially equivalent benefits (having the same average value as the defined standard benefit but with alternative benefit designs) also remained fairly stable at 40 percent. Actuarially equivalent plans continue to attract the largest share of PDP enrollees (55 percent), but the share of enrollees choosing enhanced benefit plans rose slightly to 45 percent in 2020 compared with 42 percent in 2019.
• In 2020, 69 percent of PDPs use the same $435 deductible as in Part D’s defined standard benefit, up significantly from 52 percent in 2019. Only 15 percent of PDP enrollees are in plans with no deductible. Also in 2020, 53 percent of all PDPs designate certain formulary tiers that are not subject to the deductible. If, for example, a PDP used such a designation for preferred generic drugs, an enrollee would pay just the plan’s cost sharing for that tier rather than the full cost of the prescription up to the amount of the deductible. In 2020, 56 percent of PDP enrollees were in such plans, up from 39 percent in 2019.
154 Prescription drugs
Chart 10-10. Characteristics of MA–PDs 2019 2020
Enrollees as of Enrollees as of Plans February 2019 Plans February 2020 Number Number Number Percent (in millions) Percent Number Percent (in millions) Percent
Totals 2,414 100% 13.8 100% 2,799 100% 15.3 100%
Type of organization Local HMO 1,601 66 9.7 70 1,848 66 10.6 69 Local PPO 751 31 3.3 24 891 32 4.0 26 PFFS 29 1 0.1 1 26 1 0.1 0 Regional PPO 33 1 0.8 6 34 1 0.7 4
Type of benefit Defined standard 37 2 0.1 <0.5 43 2 0.1 <0.5
Actuarially equivalent 83 3 0.2 2 81 3 0.2 2 Enhanced 2,294 95 13.5 98 2,675 96 15.0 98
Type of deductible Zero 1,116 46 6.4 46 1,349 48 7.4 49 Reduced 1,138 47 7.0 50 1,244 44 7.3 48
Defined standard* 160 7 0.5 3 206 7 0.5 4
Some formulary tiers not subject to a deductible Some 1,225 51 7.2 52 1,386 50 7.7 50 None 1,189 49 6.6 48 1,413 50 7.6 50 Note: MA–PD (Medicare Advantage–Prescription Drug [plan]), HMO (health maintenance organization), PPO (preferred
provider organization), PFFS (private fee-for-service). The MA–PDs and enrollment described here exclude employer-only
plans, plans offered in U.S. territories, 1876 cost plans, special needs plans, demonstrations, and Part B–only plans. Components may not sum to totals due to rounding. “Actuarially equivalent” includes both actuarially equivalent standard and basic alternative benefits. “Enhanced” refers to plans with basic plus supplemental coverage.
*The defined standard benefit’s deductible was $415 in 2019 and is $435 in 2020. Source: MedPAC analysis of CMS landscape, premium, and enrollment data.
• There are nearly 16 percent more MA–PDs in 2020 than in 2019. Sponsors are offering 2,799 MA–PDs in 2020 compared with 2,414 the year before. Enrollment in MA–PDs grew from 13.8 million in 2019 to 15.3 million in 2020 (10.6 percent).
• Between 2019 and 2020, the number of drug plans offered by HMOs grew from 1,601 to 1,848 and remain the dominant type of MA–PD, making up 66 percent of all offerings. Over the same period, the number of drug plans offered by local PPOs also increased from 751 plans to 891 plans.
• A larger share of MA–PDs than stand-alone prescription drug plans (PDPs) offer enhanced benefits (compare Chart 10-10 with Chart 10-9). In 2020, 60 percent of all PDPs have enhanced benefits compared with 96 percent of MA–PDs. In 2020, enhanced MA–PDs attracted 98 percent of total MA–PD enrollment.
• Forty-eight percent of MA–PDs have no deductible in 2020, and those plans attracted 49 percent of all MA–PD enrollees.
• In 2020, 50 percent of MA–PDs designated certain cost-sharing tiers of their formularies that are not subject to a deductible. Those plans account for 50 percent of MA–PD enrollment.
A Data Book: Health care spending and the Medicare program, July 2020 155
Chart 10-11. Change in average Part D premiums, 2016–2020
Average monthly premium weighted by enrollment Cumulative change
in weighted average
premium, 2016–2020
2016 2017 2018 2019 2020
All plans $31 $32 $32 $29 $27 –12 %
Basic plans 28 30 30 32 30 7
Enhanced plans
Basic benefits 27 27 26 22 20 –26
Supplemental benefits 7 6 7 6 6 –5
Total premium 33 33 33 28 26 –22
All basic coverage 27
29
28
25
23
–15
PDPs 39 41 41 40 38 –2
Basic coverage 29 31 31 32 30 4
Enhanced coverage
Basic benefits 41 43 42 35 33 –20
Supplemental benefits 12 11 15 15 15 25
Total premium 53 54 57 50 48 –10
All basic coverage 34 36 35 33 31 –8
MA–PDs, including SNPs 18 19 18 16 15 –17
Basic coverage 22 26 28 28 26 19
Enhanced coverage
Basic benefits 15 16 15 13 12 –18
Supplemental benefits 2 2 1 1 1 –43
Total premium 17 18 17 14 13 –21
All basic coverage 16 18 17 15 14 –16
MA–PD buy-down of basic premium
15 16 16 16 15 1
MA–PD buy-down of supplemental benefits
14 15 16 17 20 43
Base beneficiary premium 34.10 35.63 35.02 33.19 32.74 –4
Note: PDP (prescription drug plan), MA–PD (Medicare Advantage–Prescription Drug [plan]), SNP (special needs plan). All
calculations exclude employer-only groups and plans offered in U.S. territories. In addition, MA–PDs exclude Part B–only
plans, demonstrations, and 1876 cost plans. The MA–PD data reflect the portion of Medicare Advantage plans’ total monthly premium attributable to Part D benefits for plans that offer Part D coverage, as well as Part C rebate dollars that were used to offset Part D premium costs. The fact that average premiums for enhanced MA–PDs are lower than for basic MA–PDs could
reflect several factors such as changes in enrollment among plan sponsors and counties of operation and differences in the average health status of plan enrollees. Cumulative changes were calculated from unrounded data. Components may not sum to totals due to rounding.
Source: MedPAC analysis of CMS landscape, plan report, enrollment data, and bid data.
(Chart continued next page)
156 Prescription drugs
Chart 10-11. Change in average Part D premiums, 2016–2020 (continued)
• Part D enrollees can select between plans with basic or enhanced benefits (the latter combine basic and supplemental coverage). Medicare aims to subsidize 74.5 percent of the average cost of basic benefits; enrollees pay premiums for the remaining 25.5 percent and all of the cost of any supplemental benefits. (For more about how plan premiums are determined, see Part D Payment Basics at http://www.medpac.gov/docs/default-source/payment-basics/medpac_payment_basics_19_partd_final_sec.pdf?sfvrsn=0.)
• The overall average premium paid by enrollees for any type of Part D coverage declined from $29 per month in 2019 to $27 per month in 2020. Over the period from 2016 to 2020, year-to-year changes in average premiums have varied by type of benefit (basic vs. enhanced) and type of plan (PDP vs. MA−PD); the changes have not necessarily corresponded to changes observed in the base beneficiary premium.
• Across all basic plans and the basic portion of enhanced plans, the average premium for basic benefits fell from $27 in 2016 to $23 per month in 2020, a cumulative decline of 15 percent. This decline occurred despite very rapid growth in spending for Part D’s catastrophic phase of the benefit (data not shown). In the catastrophic phase, Medicare subsidizes 80 percent of enrollees’ drug spending. (For more information about Medicare’s Part D spending, see Chapter 14 of the Commission’s March 2020 report to the Congress at http://medpac.gov/docs/default-source/reports/mar20_medpac_ch14_sec.pdf?sfvrsn=0.)
• Over the five-year period, the average enrollee premium for basic coverage in PDPs ranged between a low of $29 in 2016 and a high of $32 per month in 2019. Between 2016 and 2020, the average premium increased by a cumulative 4 percent. Among enhanced plans offered by PDPs, the average enrollee premium has ranged from $48 in 2020 to $57 in 2018. Over the five-year period, the average premium decreased by a cumulative 10 percent. Of the $48 average premium in 2020 among enhanced PDPs, $33 was for basic benefits and $15 was for supplemental benefits. The portion of enhanced premiums attributable to supplemental benefits has grown quickly, while the portion for basic benefits has declined.
• The average Part D premium paid by beneficiaries enrolled in MA−PDs with basic coverage ranged between a low of $22 in 2016 and a high of $28 per month in 2018 and 2019. From 2016 to 2020, the average premium increased by a cumulative 19 percent. The average premium paid by beneficiaries enrolled in MA−PDs offering enhanced coverage has decreased from $17 in 2016 to $13 in 2020, a cumulative 21 percent decrease. MA−PD sponsors typically use a portion of Medicare’s Part C (Medicare Advantage) payments to “buy down” the premiums that plan enrollees would otherwise have to pay for Part D basic premiums and supplemental benefits. Because of those Part C payment “rebates,” in 2020, MA−PD enrollees avoided having to pay $15 per month in basic premiums and an additional $20 per month for supplemental coverage, on average.
A Data Book: Health care spending and the Medicare program, July 2020 157
Chart 10-12. More premium-free PDPs for LIS enrollees in 2020
Number of PDPs
Number of PDPs that have zero premium for LIS enrollees
PDP region State(s) 2019* 2020* Difference 2019* 2020 Difference
1 ME, NH 26 26 0 7 6 –1
2 CT, MA, RI, VT 26 25 –1 7 7 0
3 NY 23 27 4 8 9 1
4 NJ 26 28 2 6 8 2
5 DC, DE, MD 25 27 2 9 10 1
6 PA, WV 30 31 1 9 10 1
7 VA 27 29 2 6 7 1
8 NC 28 28 0 7 9 2
9 SC 26 28 2 3 5 2
10 GA 26 28 2 4 6 2
11 FL 27 27 0 2 4 2
12 AL, TN 29 30 1 6 7 1
13 MI 29 30 1 9 9 0
14 OH 26 28 2 7 2 –5
15 IN, KY 26 28 2 7 7 0
16 WI 28 30 2 8 9 1
17 IL 27 28 1 7 8 1
18 MO 26 28 2 4 5 1
19 AR 26 27 1 4 6 2
20 MS 24 25 1 5 7 2
21 LA 26 26 0 8 9 1
22 TX 27 30 3 5 5 0
23 OK 28 29 1 7 8 1
24 KS 26 28 2 4 6 2
25 IA, MN, MT, ND, NE, SD, WY 28 29 1
6 8 2
26 NM 27 26 –1 7 7 0
27 CO 26 26 0 7 7 0
28 AZ 28 31 3 10 12 2
29 NV 26 28 2 3 5 2
30 OR, WA 26 28 2 7 8 1
31 ID, UT 26 28 2 8 8 0
32 CA 30 32 2 7 8 1
33 HI 24 25 1 4 5 1
34 AK 22 24 2 7 7 0
Total 901 948 47 215 244 29
Note: LIS (low-income [drug] subsidy), PDP (prescription drug plan).
*These figures include two plans in 2019 that did not accept new enrollees because of CMS sanctions. Source: MedPAC based on 2019 and 2020 Part D plan report file provided by CMS.
• The total number of stand-alone PDPs increased by 5 percent, from 901 in 2019 to 948 in 2020. The median number of plans offered in PDP regions increased to 28 plans from 26 in 2019 (data not shown). In 2020, Alaska has the fewest stand-alone PDPs, with 24, and Region 32 (California) had the most, with 32.
• In 2020, 244 PDPs qualify as premium free to LIS enrollees. With the exception of Ohio, which has only two plans with no premium for LIS enrollees, at least four premium-free PDPs are available in any given region.
158 Prescription drugs
Chart 10-13. In 2020, most Part D enrollees are in plans that use a five-tier formulary structure
Note: PDP (prescription drug plan), MA–PD (Medicare Advantage–Prescription Drug [plan]). Calculations are weighted by enrollment. All calculations exclude employer-only groups and plans offered in U.S. territories. In addition, MA–PDs exclude demonstration programs, special needs plans, and 1876 cost plans. Components may not sum to totals due to
rounding. All stand-alone PDP enrollees and about 98 percent of MA−PD enrollees have a specialty tier in addition to the tiers listed above.
Source: MedPAC analysis of formularies submitted to CMS.
• Most Part D enrollees choose plans that have a five-tier structure: two generic, one preferred brand-
name tier, and one nonpreferred drug tier (which may include both brand-name and generic drugs), plus a specialty tier. In 2020, nearly all PDP enrollees continue to enroll in plans with this five-tier structure. Seventy-nine percent of MA–PD enrollees are in such plans in 2020, a slight decrease from 81 percent in 2019.
• For enrollees in PDPs with a five-tier structure, the median copay in 2020 is $43 for a preferred brand-name drug and 38 percent coinsurance for a nonpreferred drug (data not shown). The median copay for a generic drug is $0 for drugs on a lower tier and $4 for those on a higher tier. For MA–PD enrollees, in 2020, the median copay is $47 for a preferred brand and $100 for a nonpreferred brand. The median copays for generic drugs are $2 and $10 for the two generic tiers, respectively.
• All stand-alone PDPs and about 98 percent of MA–PDs use a specialty tier for drugs that have a negotiated price of $670 per month or more. In 2020, median cost sharing for a specialty-tier drug is 25 percent among PDPs and 31 percent among MA–PDs (data not shown).
4%
20%
96%
79%
0.6%
0.4%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2007 2017 2018 2019 2020 2007 2017 2018 2019 2020
25% coinsurance
Three tiers: Generic, brand, and specialty
Four tiers: Generic, preferred brand, nonpreferred, and specialty
Five tiers: Preferred generic, other generic, preferred brand, nonpreferred, and specialty
Other tier structure
PDP enrollees MA–PD enrollees
A Data Book: Health care spending and the Medicare program, July 2020 159
Chart 10-14. In 2020, the share of listed drugs subject to some utilization management increased slightly
Note: PDP (prescription drug plan), MA–PD (Medicare Advantage–Prescription Drug [plan]). Calculations are weighted by
enrollment. All calculations exclude employer-only groups and plans offered in U.S. territories. In addition, MA–PDs exclude demonstration programs, special needs plans, and 1876 cost plans. Values reflect the share of listed chemical entities that are subject to utilization management, weighted by plan enrollment. “Prior authorization” means that the
enrollee must get preapproval from the plan before coverage. “Step therapy” refers to a requirement that the enrollee try specified drugs before being prescribed other drugs in the same therapeutic category. “Quantity limits” means that plans limit the number of doses of a drug available to the enrollee in a given time period.
Source: MedPAC analysis of formularies submitted to CMS.
• In addition to the number of drugs listed on a plan’s formulary, plans’ processes for nonformulary exceptions and use of utilization management tools—prior authorization (preapproval for coverage), quantity limits (limitations on the number of doses of a particular drug covered in a given period), and step therapy requirements (enrollees being required to try specified drugs before being prescribed other drugs in the same therapeutic category)—can affect access to certain drugs.
• In 2020, the use of some form of utilization management, on average, increased slightly to 47 percent of drugs listed on a plan’s formulary in stand-alone PDPs and 46 percent in MA–PDs. Part D plans typically use quantity limits or prior authorization to manage enrollees’ prescription drug use.
• Among the drugs listed on plan formularies, on average, the share that requires prior authorization in 2020 increased to just over a quarter for stand-alone PDPs while the share for MA–PDs remained the same at 24 percent. The share with quantity limits increased for both types of plans. In 2020, on average, quantity limits apply to 33 percent of drugs listed on formularies of both stand-alone PDPs and MA–PDs. The share of drugs listed on plan formularies that requires the use of step therapy remained very low for both stand-alone PDPs and MA–PDs.
0
10
20
30
40
50
60
Priorauthorization
Steptherapy
Quantitylimits
Anyutilization
management
Perc
en
t o
f liste
d d
rug
s
2007 2017 2018 2019 2020
PDPs
0
10
20
30
40
50
60
Priorauthorization
Steptherapy
Quantitylimits
Anyutilization
management
2007 2017 2018 2019 2020
MA–PDs
160 Prescription drugs
Chart 10-15. Characteristics of Part D enrollees, 2018
All Medicare Part D
Plan type Subsidy status
PDP MA–PD LIS Non-LIS
Beneficiaries* (in millions) 62.8 46.8 27.2 19.5 14.0 32.8 Percent of all Medicare 100% 74% 43% 31% 22% 52%
Gender Male 46% 43% 43% 43% 41% 44% Female 54 57 57 57 59 56
Race/ethnicity White, non-Hispanic 74 73 78 66 53 81 African American,
non-Hispanic 10 11 9 13 20 7
Hispanic 9 10 6 15 17 7 Asian 3 4 3 4 6 2 Other 3 3 3 2 3 3
Age (years)** <65 17 17 18 16 40 7 65–69 27 25 25 25 18 27 70–74 21 22 21 23 13 25 75–79 15 16 15 16 10 18 80+ 20 21 22 20 18 22
Note: PDP (prescription drug plan), MA–PD (Medicare Advantage–Prescription Drug [plan]), LIS (low-income [drug] subsidy).
Components may not sum to totals due to rounding. *Figures for “All Medicare” and “Part D” include all beneficiaries with at least one month of enrollment in the respective
program. A beneficiary was classified as “LIS” if that individual received Part D’s LIS at some point during the year. For
individuals who switched plan types during the year, classification into plan types was based on the greater number of months of enrollment.
**Age as of July 2018.
Source: MedPAC analysis of Medicare Part D denominator file from CMS.
• In 2018, nearly 47 million Medicare beneficiaries (74 percent) were enrolled in Part D at some point in the year. About 27 million were in stand-alone PDPs, and the remaining 19.5 million were in MA–PDs. Fourteen million enrollees received Part D’s LIS.
• Demographic characteristics of Part D enrollees are generally similar to the overall Medicare population, with the exception of gender (Part D enrollees are more likely to be female). MA–PD enrollees are less likely to be disabled beneficiaries under age 65 and more likely to be Hispanic or African American compared with PDP enrollees; LIS enrollees are more likely to be female, minority, and disabled beneficiaries under age 65 compared with non-LIS enrollees.
A Data Book: Health care spending and the Medicare program, July 2020 161
Chart 10-16. Part D enrollment trends, 2007–2018
Average annual growth rate
2007 2010 2014 2018
2007− 2010
2010− 2014
2014− 2018
Part D enrollment (in millions)* Total 26.1 29.7 40.0 46.8 4.4 % 7.7 % 4.0 % Employer group waiver plans 2.0 2.6 7.0 7.3 9.2 27.4 0.8
By plan type
PDP 18.3 18.9 25.1 27.2 1.1 7.3 2.1 MA–PD 7.8 10.6 14.9 19.5 10.9 8.9 6.9
By subsidy status
LIS 10.4 11.3 12.8 14.0 2.7 3.1 2.2 Non-LIS 15.7 18.4 27.2 32.8 5.5 10.2 4.8
By race/ethnicity
White, non-Hispanic 19.4 22.0 29.6 34.1 4.3 7.7 3.6 African American, non-Hispanic 2.9 3.3 4.4 5.1 4.1 7.4 3.6 Hispanic 2.5 3.0 3.9 4.7 5.8 6.7 4.8 Other 1.3 1.4 2.1 2.9 3.9 10.3 8.8
By age (years)**
<65 5.5 6.3 7.8 8.0 4.7 5.5 0.8 65–69 5.4 6.6 9.5 11.6 6.5 9.9 4.9 70–79 8.8 9.9 13.9 17.4 3.8 8.9 5.8 80+ 6.4 7.1 8.8 9.8 3.2 5.7 2.7
Part D enrollment (in percent) Total 100 % 100 % 100 % 100 % Employer group waiver plans 8 9 17 16
By plan type
PDP 70 64 63 58 MA–PD 30 36 37 42
By subsidy status
LIS 40 38 32 30 Non-LIS 60 62 68 70
By race/ethnicity
White, non-Hispanic 74 74 74 73 African American, non-Hispanic 11 11 11 11 Hispanic 10 10 10 10 Other 5 5 5 6
By age (years)**
<65 21 21 19 17 65–69 21 22 24 25 70–79 34 33 35 37 80+ 25 24 22 21
Note: PDP (prescription drug plan), MA–PD (Medicare Advantage–Prescription Drug [plan]), LIS (low-income [drug] subsidy). A
beneficiary was classified as “LIS” if that individual received Part D’s LIS at some point during the year. If a beneficiary was enrolled in both a PDP and an MA–PD during the year, that individual was classified into the type of plan with the greater number of months of enrollment. Components may not sum to totals due to rounding. Average annual growth rate is
calculated on unrounded numbers. *Figures include all beneficiaries with at least one month of enrollment. **Age as of July of the respective year.
Source: MedPAC analysis of Medicare Part D denominator and common Medicare environment files from CMS.
(Chart continued next page)
162 Prescription drugs
Chart 10-16. Part D enrollment trends, 2007–2018 (continued)
• Part D enrollment grew faster between 2010 and 2014 (average annual growth rate (AAGR) of 7.7 percent) than between 2007 and 2010 (AAGR of 4.4 percent) or between 2014 and 2018 (AAGR of 4.0 percent). The faster enrollment growth between 2010 and 2014 largely reflects the growth in enrollment in Part D plans operated by employers for their retirees (employer group waiver plans, or EGWPs). Enrollment in EGWPs grew from 2.6 million to 7.0 million (AAGR of 27.4 percent) during this period.
• The number of enrollees receiving the LIS grew modestly between 2007 and 2018, with an AAGR of between 2.2 percent (from 2014 to 2018) and 3.1 percent (from 2010 to 2014). During the same period, the number of non-LIS enrollees grew faster than LIS enrollees, with an AAGR of 10.2 percent between 2010 and 2014 and an AAGR of 4.8 percent or greater before 2010 and after 2014. Faster enrollment growth among non-LIS enrollees is partly attributable to the recent growth in EGWPs that shifted beneficiaries into Part D plans from employer plans that had previously received Medicare’s retiree drug subsidy (RDS) (see Chart 10-7 for information on the RDS).
• Between 2014 and 2018, the largest growth in enrollment was observed for beneficiaries ages 70 to 79 (5.8 percent annually, on average), reflecting the aging of the baby boom cohort.
• While MA–PD enrollment growth decelerated in recent years from the nearly 11 percent AAGR observed between 2007 and 2010, enrollment in MA–PDs continued to exceed that of PDPs between 2014 and 2018 (AAGR of 6.9 percent and 2.1 percent, respectively).
A Data Book: Health care spending and the Medicare program, July 2020 163
Chart 10-17. Part D enrollment by region, 2018
Share of
Medicare enrollment Share of Part D enrollment*
Plan type Subsidy status
PDP
region State(s) Part D* RDS EGWP
PDP MA–PD
LIS Non-LIS
1 ME, NH 71 % 3 % 9 % 71 % 29 % 32 % 68 %
2 CT, MA, RI, VT 78 2 15 66 34 34 66
3 NY 79 4 18 53 47 37 63
4 NJ 75 3 18 78 22 24 76
5 DE, DC, MD 65 3 15 84 16 31 69
6 PA, WV 77 3 14 56 44 28 72
7 VA 65 2 9 74 26 28 72
8 NC 75 2 12 57 43 30 70
9 SC 73 2 13 65 35 29 71
10 GA 74 2 12 52 48 34 66
11 FL 77 3 7 45 55 29 71
12 AL, TN 75 2 9 51 49 34 66
13 MI 80 3 26 70 30 25 75
14 OH 79 3 13 59 41 26 74
15 IN, KY 77 2 12 68 32 30 70
16 WI 73 2 9 56 44 24 76
17 IL 74 4 12 70 30 28 72
18 MO 76 2 9 60 40 26 74
19 AR 71 3 3 70 30 37 63
20 MS 73 1 3 76 24 44 56
21 LA 76 4 9 56 44 40 60
22 TX 73 2 11 58 42 32 68
23 OK 67 1 9 76 24 31 69
24 KS 72 1 4 80 20 23 77
25 IA, MN, MT, NE, ND, SD, WY 75
2
6
73
27
22
78
26 NM 73 2 13 55 45 39 61
27 CO 74 2 10 54 46 24 76
28 AZ 75 2 8 50 50 27 73
29 NV 70 2 6 50 50 26 74
30 OR, WA 69 5 7 51 49 27 73
31 ID, UT 71 2 7 54 46 22 78
32 CA 80 2 13 48 52 35 65
33 HI 71 2 25 38 62 26 74
34 AK 42 26 3 98 2 52 48
Mean 74 2 12 58 42 30 70
Minimum 42 1 3 38 2 22 48
Maximum 80 26 26 98 62 52 78
Note: PDP (prescription drug plan), RDS (retiree drug subsidy), EGWP (employer group waiver plans), MA–PD (Medicare Advantage–Prescription Drug [plan]), LIS (low-income [drug] subsidy). Definition of regions is based on PDP regions used in Part D. If an employer agrees to provide primary drug coverage to its retirees with a benefit value that is equal to or
greater than that of Part D, Medicare provides the employer with an RDS (see Chart 10-7). *Includes enrollment in Part D plans operated for employees and their retirees (EGWPs).
Source: MedPAC analysis of Medicare Part D denominator and common Medicare environment files from CMS.
• Among Part D regions in 2018, all but one region (Region 34 (Alaska, or AK)) had 65 percent or more of all Medicare beneficiaries enrolled in Part D. (Beneficiaries in Alaska are less likely to enroll in Part D because alternative employer-sponsored drug coverage is more widely available: The share of Medicare beneficiaries enrolled in employer-sponsored plans that received the RDS was 26 percent, compared with an average of 2 percent nationwide.) In some other regions with lower than average enrollment in Part D (Region 5 and Region 7), many beneficiaries likely received their drug coverage through the Federal Employees Health Benefits Program, which does not receive the RDS.
(Chart continued next page)
164 Prescription drugs
Chart 10-17. Part D enrollment by region, 2018 (continued)
• In 2018, all regions except Region 26 and Region 34 experienced a decrease in the number of beneficiaries who received the RDS (data not shown). Since 2010, many employers have switched from operating RDS-eligible employer plans to sponsoring Part D plans for their retirees (EGWPs). In 2018, 12 percent of Medicare beneficiaries were enrolled in EGWPs compared with 2 percent who were in employer plans that received the RDS (see Chart 10-7 for information on the RDS).
• The share of Medicare beneficiaries in EGWPs varied from 3 percent in Region 34 (AK) to about a quarter in Region 13 (MI) and Region 33 (HI).
• Wide variation was seen in the shares of Part D beneficiaries who enrolled in PDPs and MA–PDs across PDP regions. The pattern of MA–PD enrollment is generally consistent with availability of and enrollment in Medicare Advantage plans.
• The share of Part D enrollees receiving the LIS ranged from 22 percent in Region 25 (IA, MN, MT, NE, ND, SD, and WY) and Region 31 (ID and UT) to 52 percent in Region 34 (AK). In all but 2 of the 34 PDP regions, LIS enrollees accounted for 40 percent or less of total Part D enrollment.
A Data Book: Health care spending and the Medicare program, July 2020 165
Chart 10-18. Components of Part D spending growth
2009 2018
Average annual growth
2009–2018
Total gross spending (in billions) $73.7 $168.1 9.6%
High-cost beneficiaries 29.2 102.2 14.9%
Lower cost beneficiaries 44.6 65.9 4.4%
Number of beneficiaries using a Part D drug (in millions) 26.5 43.5 5.7%
High-cost beneficiaries 2.4 3.9 5.5%
Lower cost beneficiaries 24.1 39.7 5.7%
Amount per beneficiary who used Part D drugs Gross drug spending per year $2,781 $3,861 3.7% Average price per 30-day prescription $55 $69 2.5%
Number of 30-day prescriptions 50.4 55.9 1.1%
Amount per high-cost beneficiary who used Part D drugs Gross drug spending per year $12,294 $26,482 8.9% Average price per 30-day prescription $110 $247 9.3%
Number of 30-day prescriptions 111.4 107.4 –0.4%
Amount per lower cost beneficiary who used Part D drugs Gross drug spending per year $1,846 $1,662 –1.2% Average price per 30-day prescription $42 $33 –2.6%
Number of 30-day prescriptions 44.5 50.9 1.5%
Note: “High-cost beneficiaries” refers to individuals who incurred spending high enough to reach the catastrophic phase of the benefit. “Gross spending” reflects payments to pharmacies from all payers, including beneficiary cost sharing, but does
not include rebates and discounts from pharmacies and manufacturers that are not reflected in prices at the pharmacies. Changes in the average price per prescription reflect both price inflation and changes in the mix of drugs used. Components may not sum to totals due to rounding.
Source: MedPAC analysis of Part D prescription drug event data and denominator files from CMS.
• Between 2009 and 2018, gross spending on drugs under the Part D program grew by an annual average rate of 9.6 percent. The annual growth in spending was considerably higher (14.9 percent) among high-cost beneficiaries (individuals who incurred spending high enough to reach the catastrophic phase of the benefit) compared with 4.4 percent for lower cost beneficiaries.
• During the 2009 through 2018 period, the number of beneficiaries who used Part D drugs grew by an annual average rate of 5.7 percent. Similar rates of growth were observed among high-cost beneficiaries and lower cost beneficiaries.
• The average price per 30-day prescription covered under Part D rose from $55 in 2009 to $69 in 2018. Overall, growth in price per prescription accounted for nearly two-thirds (2.5 percentage points) of the 3.7 percent average annual growth in spending per beneficiary among beneficiaries who used Part D drugs.
• The average annual growth rate in overall spending per beneficiary reflects two distinct patterns of price and spending growth, one for high-cost beneficiaries and another for lower cost beneficiaries. Among high-cost beneficiaries, annual growth in prices (9.3 percent) accounted for all of the spending growth (8.9 percent) during this period. In contrast, among lower cost beneficiaries, the average annual decrease in prices (–2.6 percent) resulted in an overall decrease in spending (–1.2 percent annually), despite an increase in the number of prescriptions filled during the same period.
166 Prescription drugs
Chart 10-19. The majority of Part D spending was incurred by just over one-fifth of all Part D enrollees, 2018
Note: “Spending” (gross) reflects payments from all payers, including beneficiaries (cost sharing) but does not include rebates
and discounts from pharmacies and manufacturers that are not reflected in prices at the pharmacies. In 2018, the defined standard basic benefit included a $405 deductible and 25 percent coinsurance until the enrollee reached $3,750 in total
covered drug spending. An individual with an average mix of drugs who did not receive Part D’s low-income subsidy and who had no other supplemental coverage would have reached the catastrophic phase of the benefit at about $8,418 in total drug spending. In 2018, among those who reached the catastrophic phase of the benefit, an enrollee at the 75th
percentile of the distribution had drug spending totaling $26,462. Components may not sum to totals due to rounding. Source: MedPAC analysis of Medicare Part D prescription drug event data from CMS.
• Medicare Part D spending is concentrated in a subset of beneficiaries. In 2018, about 21 percent of
Part D enrollees had annual spending exceeding the initial coverage limit (typically set at $3,750 in gross drug spending), at which point enrollees were responsible for a higher proportion of the cost of the drugs until they reached the catastrophic phase of the benefit (at about $8,418 in gross drug spending under the defined standard benefit for beneficiaries not receiving Part D’s low-income subsidy (LIS)). These beneficiaries accounted for 83 percent of total Part D spending.
• The costliest 8 percent of beneficiaries, those with drug spending above the catastrophic threshold, accounted for about 61 percent of total Part D spending. Seventy percent of beneficiaries with the highest spending received the LIS (data not shown; see Chart 10-20). Spending on prescription drugs has become more concentrated over time. Before 2011, the costliest 8 percent of beneficiaries accounted for 40 percent or less of total Part D spending (data not shown). In comparison, for Medicare Part A and Part B spending, Medicare fee-for-service spending accounted for by the costliest 5 percent of beneficiaries has been stable at about 40 percent for many years (data not shown; see Chart 1-11 for 2017 figures).
• In 2018, among Part D enrollees who reached the catastrophic phase of the benefit, those enrollees with annual spending at or above $26,462 (2 percent of all Part D enrollees) accounted for 37 percent of total Part D spending.
7
11
1
60
16
13
22
6
24
2
37
0
10
20
30
40
50
60
70
80
90
100
Percent of enrollees Percent of spending
Perc
en
t
Catastrophic phase (≥$26,462)
Catastrophic phase(<$26,462)
Coverage gapphase
Initial coveragephase
Deductible
$0
83
Total annual drug spending, by benefit
phase reached in 2018
A Data Book: Health care spending and the Medicare program, July 2020 167
Chart 10-20. Characteristics of Part D enrollees, by benefit phase reached, 2018
Annual drug spending
Below initial
coverage limit Coverage-gap
phase Catastrophic phase
Sex Male 43% 43% 43% Female 57 57 57
Race/ethnicity White, non-Hispanic 73 75 66 African American, non-Hispanic 10 10 15 Hispanic 10 9 12 Other 6 5 7
Age (years) <65 15 16 37 65–69 26 20 19 70–74 22 21 17 75–80 16 18 12 80+ 21 26 15
LIS status* LIS 25 32 70 Non-LIS 75 68 30
Plan type** PDP 57 62 65 MA–PD 43 38 35
Note: LIS (low-income [drug] subsidy), PDP (prescription drug plan), MA−PD (Medicare Advantage−Prescription Drug [plan]). “Spending” (gross) reflects payments from all payers, including beneficiaries (cost sharing) but does not include rebates
and discounts from pharmacies and manufacturers that are not reflected in prices at the pharmacies. In 2018, the defined standard basic benefit included a $405 deductible and 25 percent coinsurance until the enrollee reached $3,750 in total covered drug spending. An individual with an average mix of drugs who did not receive Part D’s low-income subsidy and
who had no other supplemental coverage would have reached the catastrophic phase of the benefit at about $8,418 in total drug spending. A small number of beneficiaries were excluded from the analysis because of missing data. Components may not sum to 100 due to rounding.
*A beneficiary was assigned LIS status if that individual received Part D’s LIS at some point during the year. **If a beneficiary was enrolled in both a PDP and an MA–PD during the year, that individual was classified in the type of
plan with the greater number of months of enrollment.
Source: MedPAC analysis of Medicare Part D prescription drug event data and Part D denominator file from CMS.
• In 2018, Part D enrollees who reached the catastrophic phase of the benefit were more likely to be minority, disabled and under age 65, and receiving the LIS compared with Part D enrollees with annual spending below the catastrophic threshold.
• While LIS enrollees are more likely to reach the catastrophic phase of the benefit, their share has been declining, from more than 80 percent in 2010 and earlier years (data not shown) to 70 percent in 2018. This decline reflects more rapid growth in enrollment of individuals who do not receive the LIS as well as the growth in average prices of drugs taken by those individuals.
• Part D enrollees who reached the catastrophic phase of the benefit were more likely to be enrolled in stand-alone PDPs (65 percent) compared with enrollees whose spending was below the initial coverage limit (57 percent) or enrollees in the coverage gap who did not reach the catastrophic threshold (62 percent). Some of this difference likely reflects the facts that LIS enrollees are more costly on average and are more likely to be in PDPs.
168 Prescription drugs
Chart 10-21. Part D spending and use per enrollee, 2018
Part D
Plan type LIS status
PDP MA–PD LIS Non-LIS
Total gross spending (billions)* $168.1 $106.2 $61.9 $81.8 $86.3 Total number of prescriptions
(millions) 2,433 1,428 1,005 877 1,556
Average spending per prescription $69 $74 $62 $93 $55 Per enrollee per month Total spending $317 $346 $276 $529 $229 OOP spending 33 36 28 6 44 Manufacturer gap discount 13 15 10 N/A 18 Plan liability 209 226 186 363 146 Low-income cost-sharing subsidy 47 51 40 160 N/A Other** 15 18 11 <1 21 Number of prescriptions 4.6 4.7 4.5 5.7 4.1
Note: PDP (prescription drug plan), MA–PD (Medicare Advantage–Prescription Drug [plan]), LIS (low-income [drug] subsidy), OOP
(out-of-pocket), N/A (not applicable). “Total gross spending” reflects payments from all payers, including beneficiaries (cost sharing) but does not include rebates and discounts from pharmacies and manufacturers that are not reflected in prices at the pharmacies. Part D prescription drug event (PDE) records are classified into plan types based on the contract identification on
each record. For purposes of classifying the PDE records by LIS status, monthly LIS eligibility information in Part D’s denominator file was used. Estimates are sensitive to the method used to classify PDE records to each plan type and LIS status. “Plan liability” includes plan payments for drugs covered by both basic and supplemental (enhanced) benefits. In
addition to the major categories shown in the chart, total spending includes amounts paid by other relatively minor payers such as group health plans, workers’ compensation, and charities. “Number of prescriptions” is standardized to a 30-day supply.
*”Total gross spending” includes nearly $6.9 billion in manufacturer discounts for brand-name drugs filled by non-LIS enrollees during the coverage gap.
**”Other” amount includes payments by patient assistance organizations and third-party payers other than Part D plans that
reduce the patient cost-sharing liability. Source: MedPAC analysis of Medicare Part D PDE data and Part D denominator file from CMS.
• In 2018, gross spending on drugs for the Part D program totaled $168.1 billion, with more than 60 percent ($106.2 billion) accounted for by Medicare beneficiaries enrolled in stand-alone PDPs. Part D enrollees receiving the LIS accounted for nearly half ($81.8 billion) of the total. Manufacturer discounts for brand-name drugs filled by non-LIS enrollees while they were in the coverage gap accounted for 4.1 percent of the total, or 8 percent of the gross spending by non-LIS enrollees (up from 3.8 percent and 7.5 percent, respectively, in 2017) (data not shown).
• The number of prescriptions filled by Part D enrollees totaled over 2.4 billion, with nearly 60 percent (about 1.4 billion) accounted for by PDP enrollees. The 30 percent of enrollees who received the LIS accounted for about 36 percent (877 million) of the total number of prescriptions filled.
• In 2018, Part D enrollees filled 4.6 prescriptions at $317 per month on average, an increase from $302 per month (for 4.5 prescriptions) in 2017 (2017 data not shown). The average monthly plan liability for PDP enrollees ($226) was considerably higher than that of MA–PD enrollees ($186), while the difference in average monthly OOP spending was smaller for MA–PD enrollees than PDP enrollees ($36 vs. $28, respectively). The average monthly low-income cost-sharing subsidy was higher for PDP enrollees ($51) compared with MA–PD enrollees ($40).
• Average monthly spending per LIS enrollee ($529) was more than double that of a non-LIS enrollee ($229), and the
average number of prescriptions filled per month by an LIS enrollee was 5.7 compared with 4.1 for a non-LIS enrollee. LIS enrollees had much lower monthly OOP spending, on average, than non-LIS enrollees ($6 vs. $44, respectively). Part D’s LIS pays for most of the cost sharing for LIS enrollees, averaging $160 per month in 2018.
A Data Book: Health care spending and the Medicare program, July 2020 169
Chart 10-22. Trends in Part D spending and use per enrollee per month, 2007–2018
Note: PDP (prescription drug plan), LIS (low-income [drug] subsidy), MA–PD (Medicare Advantage–Prescription Drug [plan]). “Spending” (gross) reflects payments from all payers, including beneficiaries (cost sharing) but does not include rebates and discounts from pharmacies and manufacturers that are not reflected in prices at the pharmacies. Part D prescription
drug event (PDE) records are classified into plan types based on the contract identification on each record. For purposes of classifying the PDE records by LIS status, monthly LIS eligibility information in Part D’s denominator file was used. Figures are sensitive to the method used to classify PDE records to each plan type and LIS status.
Source: MedPAC analysis of Medicare Part D PDE data and Part D denominator file from CMS.
• Between 2007 and 2018, average per capita spending per month for Part D–covered drugs grew from
$212 to $317, an average growth of 3.7 percent annually, or about 49 percent cumulatively. The rate of growth in average per capita spending more than doubled after 2013, in part reflecting the introduction of new hepatitis C treatments in 2014 and subsequent years.
• Between 2007 and 2018, monthly per capita spending for LIS enrollees grew faster than that for non-LIS enrollees, increasing from $301 to $529 (a cumulative growth of nearly 76 percent) compared with an increase from $156 to $229 for non-LIS enrollees (a cumulative growth of 47 percent). The number of prescriptions filled by both LIS and non-LIS enrollees grew by just under 2 percent annually during this period (data not shown).
• The growth in monthly per capita drug spending among MA−PD enrollees exceeded that of PDP enrollees during the 2007 to 2018 period (annual average growth of 5.6 percent and 3.4 percent, respectively). However, the average per capita spending for MA−PD enrollees continued to be lower than that of PDP enrollees (by about $70 per month in 2018).
$212
$242
$317
$239
$275
$346
$151
$185
$276
$301
$377
$529
$156$179
$229
$0
$100
$200
$300
$400
$500
$600
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Gro
ss P
art
D s
pen
din
g p
er
en
roll
ee p
er
mo
nth
All Part D PDP
MA-PD LIS
Non-LIS
170 Prescription drugs
Chart 10-23. Top 15 therapeutic classes of drugs covered under Part D, by spending and volume, 2018
Top 15 therapeutic classes by spending Top 15 therapeutic classes by volume
Dollars Prescriptions
Billions Percent Millions Percent
Diabetic therapy $26.8 15.9% Antihyperlipidemics 257.4 10.6% Asthma/COPD therapy agents 12.4 7.4 Antihypertensive therapy
agents 252.6 10.4
Antivirals 9.8 5.8 Diabetic therapy 163.9 6.7 Antineoplastic (enzyme
inhibitors) 9.5 5.6 Antidepressants 151.2 6.2
Anticoagulants 9.4 5.6 Beta-adrenergic blockers 150.4 6.2 Analgesics (anti-inflammatory/ antipyretic, non-narcotic)
7.7 4.6 Peptic ulcer therapy 123.2 5.1
Anticonvulsants 6.2 3.7 Diuretics 115.4 4.7 Antipsychotics 5.9 3.5 Calcium channel blockers 110.5 4.5 Antihypertensive therapy agents
5.0 3.0 Thyroid therapy 96.7 4.0
Antineoplastics (immunomodulators)
5.0 3.0
Anticonvulsants 94.9 3.9
Antihyperlipidemics 4.6 2.7 Analgesics (narcotic) 71.7 2.9 Analgesics (narcotic) 3.1 1.9 Asthma/COPD therapy agents 69.1 2.8 Antineoplastics (hormone antagonists)
3.0 1.8 Antibacterial agents 57.7 2.4
Antidepressants 2.7 1.6 Prostatic hypertrophy agents 48.4 2.0 Peptic ulcer therapy 2.7 1.6 Anticoagulants 45.0 1.8 Subtotal, top 15 classes 113.9 67.8 Subtotal, top 15 classes 1,808.0 74.3 Total, all classes 168.1 100.0 Total, all classes 2,432.5 100.0
Note: COPD (chronic obstructive pulmonary disease). “Spending” (gross) reflects payments from all payers, including beneficiaries (cost sharing) but does not include rebates and discounts from pharmacies and manufacturers that are not
reflected in prices at the pharmacies. “Volume” is the number of prescriptions, standardized to a 30-day supply. Therapeutic classification is based on the First DataBank Enhanced Therapeutic Classification System 1.0. Components may not sum to totals due to rounding.
Source: MedPAC analysis of Medicare Part D prescription drug event data from CMS.
• In 2018, the top 15 therapeutic classes by spending accounted for more than two-thirds of the $168.1 billion spent on prescription drugs covered by Part D plans. The top 15 therapeutic classes by volume accounted for nearly three-quarters of the over 2.4 billion prescriptions dispensed in 2018.
• While many of the same therapeutic classes on the top-15 list appear year after year, the ranking has changed from time to time. For example, market entries of new hepatitis C therapies more than tripled Part D spending on antivirals between 2013 and 2015 (data not shown). In 2018, antivirals accounted for $9.8 billion, down from $11.7 billion in 2016 (2016 data not shown). The growth in spending for drugs to treat cancer resulted in three classes of antineoplastic therapies (enzyme inhibitors, immunomodulators, and hormone antagonists) appearing on the top-15 list for the first time in 2018, compared with just one class between 2012 and 2014 and none before 2012.
(Chart continued next page)
A Data Book: Health care spending and the Medicare program, July 2020 171
Chart 10-23. Top 15 therapeutic classes of drugs covered under Part D, by spending and volume, 2018 (continued)
• Spending on drugs to treat diabetes has grown at a double-digit rate since 2007 (data not shown). In 2018, spending on diabetic therapy totaled $26.8 billion, an increase of about 15 percent from $23.3 billion in 2017 (2017 data not shown). The number of prescriptions filled for diabetic therapy totaled 163.9 million, an increase of 5.5 percent from 155.4 million in 2017.
• Nine therapeutic classes are among the top 15 in both spending and volume. Diabetic therapy dominates the list by spending, accounting for almost 16 percent of total spending and nearly a quarter of spending for the top 15 therapeutic classes, followed by asthma/COPD therapy agents. Cardiovascular agents (antihyperlipidemics, antihypertensive therapy agents, beta-adrenergic blockers, diuretics, and calcium channel blockers) dominate the list by volume, accounting for about 36 percent of all prescriptions and 50 percent of the prescriptions in the top 15 therapeutic classes.
172 Prescription drugs
Chart 10-24. Part D patterns of prescribing by provider type, 2017
Part D
Provider type
Primary care*
Specialty/ others
NP/PA/ CNS
Number of individual prescribers (thousands) 1,163 254 660 249 Share of all individual prescribers 22% 57% 21% Average beneficiary count 158 254 125 146 Average per beneficiary Gross spending $753 $912 $745 $617 Number of prescriptions 6.0 11.2 4.2 5.4 Top 1 percent of prescribers based on number of prescriptions filled per beneficiary
Number of individual prescribers 10,311 7,228 1,921 1,162 Share of top 1 percent of prescribers 70% 19% 11% Total gross spending (billions) $9.9 $7.7 $1.5 $0.7 Share of provider type’s total gross spending 6% 13% 2% 3% Total number of prescriptions (millions) 142 118 17 8 Share of provider type’s total prescriptions filled 10% 14% 4% 3%
Average per beneficiary Gross spending $3,812 $3,243 $5,371 $4,773 Number of prescriptions 42 42 42 41
Note: NP (nurse practitioner), PA (physician assistant), CNS (clinical nurse specialist). “Gross spending” reflects payments from all payers, including beneficiaries (cost sharing) but does not include rebates and discounts from pharmacies and manufacturers that are not reflected in prices at the pharmacies. Numbers may not sum to totals due to rounding.
“Number of prescriptions” is a count of prescription drug events and is not adjusted for the size (number of days’ supply) of the prescriptions. As such, these figures are not comparable with the prescription counts shown in Chart 10-18, Chart 10-21, and Chart 10-23.
*The definition of “primary care” used here includes practitioners who have a primary Medicare specialty designation of family practice, internal medicine, pediatrics, or geriatrics.
Source: MedPAC analysis of Medicare Part D prescriber-level public use file from CMS.
• In 2017, nearly 1.2 million individual providers wrote prescriptions for Medicare beneficiaries that were filled under Part D. Of those, about 22 percent were primary care providers, 57 percent were specialty or other types of providers, and 21 percent were NPs, PAs, or CNSs in primary and specialty care. While historically, NPs and PAs have been concentrated in primary care, more recent patterns suggest that they are increasingly practicing in specialty fields.
• The average count of Medicare-only beneficiaries was higher among primary care providers compared with specialty and other types of providers and with NPs, PAs, and CNSs—254 beneficiaries versus 125 beneficiaries and 146 beneficiaries, respectively.
(Chart continued next page)
A Data Book: Health care spending and the Medicare program, July 2020 173
Chart 10-24. Part D patterns of prescribing by provider type, 2017 (continued)
• On a per beneficiary basis, average gross spending for Part D prescriptions was much higher for prescriptions written by primary care providers ($912) compared with the average for specialty and other providers ($745) and for NPs, PAs, and CNSs ($617). Primary care providers also wrote more prescriptions per beneficiary, on average: 11.2 compared with 4.2 for specialty and other providers and 5.4 for NPs, PAs, and CNSs.
• More than 10,300 prescribers were among the top 1 percent of all prescribers, as ranked by the average number of Part D prescriptions filled per beneficiary in 2017. The top prescribers were much more likely than all providers to be practicing in primary care: 70 percent were primary care providers, 19 percent were specialty and other providers, and 11 percent were NPs, PAs, and CNSs.
• The top 1 percent of prescribers accounted for 6 percent of total gross spending and 10 percent of all prescriptions filled. Among primary care prescribers who were within the top 1 percent, results were more concentrated: They accounted for 13 percent of gross prescription spending and 14 percent of all prescriptions written by primary care providers.
• Among the prescriptions that were written by prescribers in the top 1 percent of all prescribers in 2017, per beneficiary Part D spending averaged $3,812 for 42 prescriptions filled.
174 Prescription drugs
Chart 10-25. Part D patterns of prescribing for selected specialties, 2017
Number of individual Part D
prescribers (thousands)
Share of all Part D
prescribers (percent)
Average per beneficiary
Gross spending (in dollars)
Number of prescriptions
All Part D 1,162.9 100 % $753 6.0
All specialty/others 659.6 57 745 4.2
Selected specialties:
Psychiatry 25.4 4 1,260 13.3
Cardiology 20.3 3 799 8.3
Ophthalmology 19.8 3 454 4.1
Psychiatry & neurology 14.2 2 1,232 11.3
Neurology 13.9 2 3,050 7.4
Gastroenterology 13.6 2 1,669 3.6
Urology 10.7 2 423 3.9
Pulmonary disease 9.5 1 2,977 6.8
Nephrology 8.6 1 1,793 8.5
Hematology & oncology 8.5 1 8,081 6.1
Endocrinology 5.9 1 2,421 8.1
Infectious disease 5.4 1 6,635 8.9
Rheumatology 4.7 1 3,374 7.9
Medical oncology 3.2 <0.5 7,422 5.7
Note: “Gross spending” reflects payments from all payers, including beneficiaries (cost sharing) but does not include rebates
and discounts from pharmacies and manufacturers that are not reflected in prices at the pharmacies. “Number of prescriptions” is a count of prescription drug events and is not adjusted for the size (number of days’ supply)
of the prescriptions. As such, they are not comparable with the prescription counts shown in Chart 10-18, Chart 10-21, and Chart 10-23.
Source: MedPAC analysis of Medicare Part D prescriber-level public use file from CMS.
• Of specialty care prescribers, psychiatrists were among the most numerous, making up 4 percent of all Part D prescribers in 2017. Cardiologists, ophthalmologists, psychiatrist/neurologists, neurologists, gastroenterologists, and urologists each made up another 2 percent to 3 percent of Part D prescribers.
• Psychiatrists wrote an average of 13.3 prescriptions per beneficiary, with an average of $1,260 in gross spending per patient. Those are higher than the overall Part D averages of 6.0 prescriptions and $753 in average gross spending per beneficiary. Other specialties with comparatively high average gross spending per beneficiary include psychiatry/neurology, neurology, gastroenterology, pulmonary disease, nephrology, hematology/oncology, endocrinology, infectious disease, rheumatology, and medical oncology.
(Chart continued next page)
A Data Book: Health care spending and the Medicare program, July 2020 175
Chart 10-25. Part D patterns of prescribing for selected specialties, 2017 (continued)
• Other specialties such as ophthalmology and urology had lower average gross spending per beneficiary. Cardiologists had average gross spending per beneficiary slightly higher than that of all Part D specialty prescribers ($799 vs. $753, respectively), but wrote an average of 8.3 prescriptions per beneficiary—considerably more than the average of 4.2 per beneficiary for all Part D specialty prescribers.
176 Prescription drugs
Chart 10-26. Price growth for Part D–covered drugs, 2006–2018
Note: Q1 (first quarter), Q4 (fourth quarter). Part D indexes reflect total amounts paid to pharmacies and do not reflect
retrospective rebates or discounts from manufacturers and pharmacies. These measures of price growth reflect growth in
the price of individual products but do not reflect changes in price due to the introduction of new products or to changes in the mix of products used.
Source: Acumen LLC analysis for MedPAC.
• Measured by individual national drug codes, prices of drugs and biologics covered under Part D rose
90 percent cumulatively between 2006 and 2018 (an index of 1.90). (Prices reflect total amounts paid to pharmacies and do not reflect retrospective rebates or discounts from manufacturers and pharmacies.)
• As measured by a price index that takes generic substitution into account, Part D prices increased by just 14 percent cumulatively (an index of 1.14) over the 12-year period. Before 2013, increased generic use kept overall prices stable by offsetting increases in prices of brand-name drugs. From 2013 to 2015, however, the introduction of new generics slowed, and prices for brand-name drugs grew more rapidly—as reflected by an uptick in the price index.
• Overall, between 2006 and 2018, prices of generic drugs covered under Part D decreased to 23 percent of the average price observed at the beginning of 2006. In comparison, prices of single-source, brand-name drugs (drugs with no generic substitutes) grew by a cumulative 244 percent (an index of 3.44) during the same period.
All Part D–covered drugs
1.90
Single-source brand-name drugs
3.44
Generic drugs0.23
All Part D–covered drugs accounting for generic substitution
1.14
0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.25
2.50
2.75
3.00
3.25
3.50
3.75
2006Q1
2007Q1
2008Q1
2009Q1
2010Q1
2011Q1
2012Q1
2013Q1
2014Q1
2015Q1
2016Q1
2017Q1
2018Q1
Ch
ain
-weig
hte
d F
ish
er
pri
ce i
nd
ex
2018 Q4
A Data Book: Health care spending and the Medicare program, July 2020 177
Chart 10-27. Comparison of price growth for Part B and Part D biologics, 2006–2018
Note: Q1 (first quarter), Q4 (fourth quarter). Part D indexes reflect total amounts paid to pharmacies and do not reflect retrospective
rebates or discounts from manufacturers and pharmacies. The Part B index reflects growth in the average sales price of Part
B–covered biologics over time, measured for individual biologics at the Healthcare Common Procedure Coding System billing code level. These measures of price growth reflect growth in the price of individual products but do not reflect changes in price due to the introduction of new products or the changes in the mix of products used. The Part B price index for biologics
in this chart and in Chart 10-6 are different due to the different periods of analysis. Source: Acumen LLC analysis for MedPAC.
• Measured by the change in the average sales price of individual Part B–covered biologics, the prices of Part B–covered biologics rose by an average of 54 percent cumulatively between 2006 and 2018 (an index of 1.54). Measured by individual national drug codes, prices of biologics covered under Part D rose 258 percent cumulatively during the same period (an index of 3.58). (Prices reflect total amounts paid to pharmacies and do not reflect retrospective rebates or discounts from manufacturers and pharmacies).
• Prices of noninsulin biologics covered under Part D grew less rapidly (by an average of 220 percent cumulatively, an index of 3.20) compared with the growth in prices of all Part D biologics during the same period.
• These measures of price growth reflect growth in price at the individual product level and do not reflect changes in price that occur as a result of shifts in the mix of biologics used or the introduction of new, higher priced biologics.
• Currently, biologics that may be covered under either Part B or Part D are limited to a subset of drugs within therapeutic classes such as therapies to treat inflammatory conditions (e.g., rheumatoid arthritis) and certain types of cancer.
Part B biologics
1.54
Part D biologics
3.58
Part D biologics excluding insulin
3.20
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
2006Q1
2007Q1
2008Q1
2009Q1
2010Q1
2011Q1
2012Q1
2013Q1
2014Q1
2015Q1
2016Q1
2017Q1
2018Q1
Ch
ain
-weig
hte
d F
ish
er
pri
ce i
nd
ex
2018 Q4
11
0
A Data Book: Health care spending and the Medicare program, July 2020 181
Chart 11-1. Number of dialysis facilities is growing, and most facilities are for profit and freestanding
Average annual percent change
2018 2013–2018 2017–2018
Total number of:
Dialysis facilities 7,441 4% 5% Hemodialysis stations 130,257 4 6
Mean number of hemodialysis stations per facility 18 –0.2 1.2 Share of total facilities
Hospital based 5% –4 –6 Freestanding 95 5 6 Urban 83 5 6 Rural, micropolitan 10 2 3 Rural, adjacent to urban 4 2 2 Rural, not adjacent to urban 2 2 –2 Frontier 0.5 1 0 For profit 88 5 5
Nonprofit 12 −0.4 2
Note: “Nonprofit” includes facilities designated as either nonprofit or government. “Average annual percent change” is based on
comparing 2013, 2017, and 2018 end-of-year files. Components may not sum to totals due to rounding. Source: Compiled by MedPAC from the institutional outpatient claims files and the Dialysis Compare files from CMS.
• Between 2013 and 2018, the number of facilities has increased, on average, 4 percent per year. The average size of a facility has remained relatively constant, averaging nearly 18 dialysis treatment stations per facility (17.7 stations in 2013, 17.3 stations in 2017, and 17.5 stations in 2018).
• Since 2013, facilities’ capacity to provide care—as measured by hemodialysis treatment stations—grew 4 percent annually on average. Capacity at urban facilities grew by 4 percent per year, while capacity at rural facilities grew at a rate of 2 percent per year (data not shown).
• Since 2013, the number of freestanding and for-profit facilities increased, while hospital-based facilities decreased. Both freestanding and for-profit facilities each increased by 5 percent per year to nearly 7,050 freestanding facilities and about 6,570 for-profit facilities.
182 Other services
Chart 11-2. Medicare spending for outpatient dialysis services furnished by freestanding and hospital-based dialysis facilities, 2017 and 2018
Note: ESRD (end-stage renal disease).
Source: Compiled by MedPAC from the institutional outpatient claims files from CMS.
• In 2018, total spending for dialysis, dialysis drugs, and ESRD-related clinical laboratory tests was $12.7 billion. Medicare paid all facilities under a prospective payment system (PPS) that includes in the payment bundle certain dialysis drugs and ESRD-related clinical laboratory tests that were separately paid before 2011.
• Between 2017 and 2018, total ESRD expenditures increased by 11 percent. Nearly all of the growth in spending is due to payments for two drugs that qualified in 2018 for the ESRD PPS’s transitional drug add-on payment adjustment (TDAPA). Without the TDAPA, dialysis spending would have increased 0.5 percent, a rate similar to the growth seen between 2016 and 2017.
• Freestanding dialysis facilities treated most dialysis beneficiaries and accounted for 96 percent of expenditures in 2018.
0
2
4
6
8
10
12
14
2017 2018
Sp
en
din
g o
n d
ialy
sis
, d
ialy
sis
dru
gs,
an
d o
ther
ES
RD
-rela
ted
serv
ices
(in
bil
lio
ns o
f d
oll
ars
)
Freestanding Hospital based
$11.4
$12.7
95% 96%
5%5%
4%
A Data Book: Health care spending and the Medicare program, July 2020 183
Chart 11-3. The ESRD population is growing, and most ESRD patients undergo dialysis
2007 2013 2017
Patients Patients Patients (thousands) Percent (thousands) Percent (thousands) Percent
Total 527.2 100% 658.4 100% 746.6 100%
Dialysis 369.3 70 462.0 70 523.7 70 In-center hemodialysis 335.4 64 408.3 62 458.6 61 Home hemodialysis* 3.7 0.7 8.1 1 9.5 1 Peritoneal dialysis* 28.7 5 43.9 7 52.7 7 Unknown 1.5 0.3 1.7 0.3 2.9 0.4 Functioning graft and kidney transplant 157.9 30 196.4 30 222.8 30 Note: ESRD (end-stage renal disease). Totals may not equal sum of components due to rounding. Data include both Medicare
(fee-for-service and Medicare Advantage) and non-Medicare patients. The “functioning graft and kidney transplant” category includes patients who have a functioning graft at the start of the year in question (i.e., 2007, 2013, or 2017), or
who receive a transplant during the year in question. *Home dialysis methods.
Source: Compiled by MedPAC from the United States Renal Data System.
• Persons with ESRD require either dialysis or a kidney transplant to maintain life. The total number of ESRD patients increased by 4 percent annually between 2007 and 2017.
• In hemodialysis, a patient’s blood flows through a machine with a special filter that removes wastes and extra fluids. In peritoneal dialysis, the patient’s blood is cleansed by using the lining of his or her abdomen as a filter. Peritoneal dialysis is the most common form of home dialysis.
• Most ESRD patients undergo hemodialysis administered in a dialysis facility three times a week. Between 2007 and 2017, the total number of in-center hemodialysis patients grew by 3 percent annually, while the total number of peritoneal dialysis patients increased by about 6 percent annually. Although a smaller proportion of all dialysis patients undergo home hemodialysis, the number of these patients grew 10 percent per year during this period.
• Functioning graft patients are patients who have had a successful kidney transplant. Patients undergoing a kidney transplant may receive either a living kidney or a cadaveric kidney donation. In 2017, 28 percent of transplanted kidneys were from living donors and the remainder were from cadaver donors (data not shown).
184 Other services
Chart 11-4. Asian Americans and Hispanics are among the fastest growing segments of the ESRD population
Share Average annual of total percent change
in 2017 2012–2017
Total (N = 746,557) 100% 3% Age (years) 0–17 1 0.4 18–44 14 1 45–64 43 3 65–79 33 6 80+ 9 4 Sex Male 58 4 Female 42 3 Race/ethnicity White 62 3 African American 30 3 Native American 1 2 Asian American 6 6 Hispanic 18 4 Non-Hispanic 80 3 Unknown 2 1 Underlying cause of ESRD Diabetes 39 4 Hypertension 26 4 Glomerulonephritis 16 2 Other causes 20 3 Note: ESRD (end-stage renal disease). Totals may not equal sum of the components due to rounding. ESRD patients include
those who undergo maintenance dialysis and those who have a functioning kidney transplant. Data include both Medicare
(fee-for-service and Medicare Advantage) and non-Medicare patients. Source: Compiled by MedPAC from the United States Renal Data System.
• Among ESRD patients, nearly 42 percent are over age 65. About 62 percent are White.
• Diabetes is the most common cause of renal failure.
• The number of ESRD patients increased by 3 percent annually between 2012 and 2017. Among the fastest growing groups of patients are patients between the ages of 65 and 79, Asian Americans, and Hispanics.
A Data Book: Health care spending and the Medicare program, July 2020 185
Chart 11-5. Characteristics of Medicare fee-for-service dialysis patients, 2018
Share of all FFS dialysis patients Age (years) Under 45 10%
45–64 38 65–74 28 75–84 18 85+ 6
Sex Male 56 Female 44 Race White 47 African American 35 All other 18 Residence Urban county 83 Rural county, micropolitan 10 Rural county, adjacent to urban 5 Rural county, not adjacent to urban 2 Frontier county 1 Prescription drug coverage status Enrolled in Part D plan or other source of creditable drug coverage 89 LIS 58 Dually eligible for Medicare and Medicaid 48 Note: FFS (fee-for-service), LIS (low-income [drug] subsidy). Urban counties contain a core area with 50,000 or more people,
rural micropolitan counties contain at least one cluster of at least 10,000 and fewer than 50,000 people, rural counties adjacent to urban areas do not have a city of 10,000 people in the county, and rural counties not adjacent to urban areas
do not have a city of 10,000 people. Frontier counties are counties with six or fewer people per square mile. Totals may not sum to 100 percent due to rounding.
Source: MedPAC analysis of dialysis claims files and denominator files from CMS.
• Compared with all Medicare patients, FFS dialysis patients are disproportionately younger and African American (see Chart 2-5).
• In 2018, about 17 percent of FFS dialysis patients resided in a rural county.
• Nearly half of all dialysis patients were dually eligible for Medicare and Medicaid services.
• Nearly 90 percent of FFS dialysis patients were enrolled in Part D plans or had other sources of creditable drug coverage.
186 Other services
Chart 11-6. Aggregate margins varied by type of freestanding dialysis facility, 2018
Share of freestanding Type of facility dialysis treatments Aggregate margin All facilities 100% 2.1% Urban 88 2.8
Rural 12 −2.8 Treatment volume (quintile)
Lowest 7 −19.3
Second 12 −8.0
Third 17 −0.1 Fourth 24 4.2 Highest 39 8.7
Note: Margins include payments and costs for dialysis services commonly provided under treatment, including injectable drugs
and laboratory tests that were separately paid before 2011. Totals may not sum to 100 percent due to rounding.
Source: Compiled by MedPAC from 2018 cost reports and the 2018 institutional outpatient file from CMS.
• For 2018, the aggregate Medicare margin for dialysis-related services, including ESRD-related drugs and laboratory tests that were separately paid before 2011, was 2.1 percent.
• Between 2017 and 2018, the aggregate Medicare margin increased (from −1.1 percent to 2.1 percent) due to the profitability of the drugs paid under the transitional drug add-on payment adjustment (TDAPA) policy. Excluding the payments and costs of the drugs paid under the TDAPA (calcimimetics), we estimate that the 2018 aggregate Medicare margin
would have been about −2 percent.
• Generally, freestanding dialysis facilities’ margins vary by the size of the facility; facilities with greater treatment volume have higher margins on average. Differences in capacity and treatment volume explain some of the differences observed between the margins of urban facilities versus rural facilities. Urban facilities are larger on average than rural facilities with respect to the number of dialysis treatment stations and Medicare treatments provided. Some rural facilities have benefited from the ESRD prospective payment system’s low-volume adjustment.
A Data Book: Health care spending and the Medicare program, July 2020 187
Chart 11-7. Dialysis quality of care: Some measures show progress, others need improvement, 2012–2017
Outcome measure 2012 2016 2017
Share of in-center hemodialysis patients: Receiving adequate dialysis 97% 98% 98% Dialyzed with an AV fistula 60 62 63*
Share of peritoneal dialysis patients receiving adequate dialysis 90 93 93 Share of all dialysis patients managing anemia
Mean hemoglobin <10 g/dL 23 29 28 Mean hemoglobin 10 to <12 g/dL 69 66 67 Mean hemoglobin ≥12 g/dL 7 5 5
Share of all dialysis patients wait-listed for a kidney 17.6 15.3 13.7
Renal transplant rate per 100 patient years 3.5 3.5 3.6
Annual mortality rate per 100 patient years** 17.0 16.4 16.5
Total hospital admissions per patient year** 1.9 1.7 1.7
Hospital days per patient year** 12.0 11.4 11.3 Note: AV (arteriovenous), g/dL (grams per deciliter [of blood]), USRDS (United States Renal Data System). Totals may not sum
to 100 percent due to rounding. The rate per patient year is calculated by dividing the total number of events by the fraction of the year that patients were followed. Data on dialysis adequacy, anemia management, and fistula utilization represent the share of patients meeting CMS’s clinical performance measures. The United States Renal Data System
(USRDS) adjusts hospitalization and mortality measures by age, gender, race, and primary diagnosis of end-stage renal disease.
*Use of AV fistula as of May 2018 (data on 2017 AV fistula use not available from USRDS).
**Lower values suggest higher quality.
Source: All measures, except for share of patients receiving adequate dialysis and anemia management, compiled by MedPAC using data from the United States Renal Data System. Measure of share of patients receiving adequate dialysis and
anemia management compiled by MedPAC using data from CMS’s 100 percent institutional outpatient files.
• Quality of dialysis care is mixed. Performance has improved on some measures, but performance on others remains unchanged or has declined.
• Between 2012 and 2017, overall adjusted mortality rates decreased from 17.0 percent to 16.5 percent. During this period, the proportion of hemodialysis patients receiving adequate dialysis remained high, and there has been a modest decline in the overall rates of hospitalization.
• All hemodialysis patients require vascular access—the site on the patient’s body where blood is removed and returned during dialysis. Use of arteriovenous fistulas, considered the best type of vascular access, has modestly increased from 60 percent to 63 percent of hemodialysis patients between 2012 and 2017.
• Other measures suggest that improvements in dialysis quality are still needed. We looked at access to kidney transplantation because it is widely believed to be the best treatment option for individuals with end-stage renal disease. Between 2012 and 2017, the share of dialysis patients accepted on the kidney transplant waiting list declined from 17.6 to 13.7, and the renal transplant rate per 100 dialysis-patient years remained relatively constant at 3.6.
188 Other services
Chart 11-8. Hospice spending and use increased in 2018
Average annual change, Change, 2000 2017 2018 2000–2017 2017–2018 Medicare payments (in billions) $2.9 $17.9 $19.2 11.2% 7.4% Beneficiaries in hospice 0.534 1.493 1.551 6.2% 3.9%
(in millions) Number of hospice days for all 25.8 106.3 113.5 8.7% 6.8%
hospice beneficiaries (in millions) Average length of stay 53.5 88.1 89.6 3.0% 1.7%
among decedents (in days)
Median length of stay 17 17 18 0 days 1 day among decedents (in days)
Note: Average length of stay is calculated for decedents who were using hospice at the time of death or before death and reflects the total number of days the decedent was enrolled in the Medicare hospice benefit during his or her lifetime. Total spending, number of hospice users, number of hospice days, and average length of stay displayed in the table are
rounded; the percentage change (except for length of stay) is calculated using unrounded data. Length-of-stay data for 2017 and 2018 are based on the Medicare Beneficiary Database obtained from CMS in October 2019. Length-of-stay figures for 2017 differ from those published in the June 2019 data book because they were based on an earlier version of
the Medicare Beneficiary Database obtained from CMS. CMS has revised the hospice election information for some beneficiaries in the Medicare Beneficiary Database.
Source: MedPAC analysis of the denominator file, the Medicare Beneficiary Database, and the 100 percent hospice claims standard
analytic file from CMS.
• Total Medicare payments to hospices were about $19.2 billion in 2018, about 7 percent higher than the prior year.
• The number of Medicare beneficiaries receiving hospice services, total number of days of hospice care, and average length of stay continued to grow in 2018.
A Data Book: Health care spending and the Medicare program, July 2020 189
Chart 11-9. Hospice use increased across beneficiary groups from 2000 to 2018
Average annual Share of decedents using hospice percentage Percentage point change point change 2000 2017 2018 2000–2017 2017–2018 All 22.9% 50.0% 50.7% 1.6 0.7% FFS beneficiaries 21.5 49.0 49.7 1.6 0.7 MA beneficiaries 30.9 52.3 52.8 1.3 0.5 Dual eligibles 17.5 44.8 45.6 1.6 0.8 Non–dual eligibles 24.5 51.7 52.4 1.6 0.7 Age (years) <65 17.0 29.6 30.0 0.7 0.4 65–84 24.7 46.7 47.1 1.3 0.4 85+ 21.4 60.1 61.4 2.3 1.3 Race/ethnicity White 23.8 52.2 53.0 1.7 0.8 Minority 17.3 39.3 39.7 1.3 0.4 Gender Male 22.4 45.5 46.1 1.4 0.6 Female 23.3 54.2 55.1 1.8 0.9 Residence Urban county 24.2 51.0 51.6 1.6 0.6 Rural county, micropolitan 18.3 46.9 47.9 1.7 1.0 Rural county, adjacent to urban 17.5 46.6 47.5 1.7 0.9 Rural county, nonadjacent to urban 15.0 41.2 42.3 1.5 1.1 Frontier county 13.1 34.1 36.1 1.2 2.0 Note: FFS (fee-for-service), MA (Medicare Advantage). “Residence” refers to the beneficiary’s county of residence. Urban,
micropolitan, and rural designations are based on the urban influence codes. This chart uses the 2013 urban influence
code definition. The frontier category is defined as population density equal to or less than six people per square mile and overlaps with the beneficiary county of residence categories. Hospice use rates for 2017 and 2018 are based on the Medicare Beneficiary Database obtained from CMS in October 2019. Hospice use rates for 2017 differ from those
published in the June 2019 data book because they were based on an earlier version of the Medicare Beneficiary Database obtained from CMS. CMS has revised the hospice election information for some beneficiaries in the Medicare Beneficiary Database.
Source: MedPAC analysis of data from the denominator file and the Medicare Beneficiary Database from CMS.
• Hospice use grew across beneficiary groups in 2018, continuing the trend of a growing proportion of beneficiaries using hospice at the end of life.
• Despite this growth, hospice use continued to vary by demographic and beneficiary characteristics. Medicare decedents who were not dual eligible, who were MA enrollees, older, White, female, or living in an urban area were more likely to use hospice than their respective counterparts.
190 Other services
Chart 11-10. Number of Medicare-participating hospices has increased due to growth in for-profit hospices
2000 2016 2017 2018 All hospices 2,255 4,382 4,488 4,639 For profit 672 2,940 3,097 3,226 Nonprofit 1,324 1,275 1,230 1,248 Government 257 167 160 158 Freestanding 1,069 3,369 3,519 3,674 Hospital based 785 501 471 454 Home health based 378 487 475 466 SNF based 22 25 22 22 Urban 1,455 3,474 3,603 3,736 Rural 757 901 879 869 Note: SNF (skilled nursing facility). Numbers may not sum to totals because of missing data for a small number of providers.
The rural and urban definitions in this chart are based on updated definitions of the core-based statistical areas (which
rely on data from the 2010 census). Source: MedPAC analysis of Medicare cost reports, Provider of Services file, and the standard analytic file of hospice claims from
CMS.
• There were 4,639 Medicare-participating hospices in 2018. Almost 70 percent of them were for-profit hospices.
• The number of Medicare-participating hospices grew by roughly 150 providers between 2017 and 2018 and has doubled since 2000. For-profit hospices accounted for most of the net growth in providers between 2017 and 2018.
• Growth in the number of providers has occurred predominantly among freestanding providers. The number of hospital-based and home health–based providers declined between 2016 and 2018. The number of SNF-based providers is small and has changed little over the years. (A hospice’s status as freestanding versus hospital based, home health based, or SNF based reflects the type of cost report submitted by the provider and does not necessarily reflect the location of care.)
• The number of hospices located in rural areas has declined in the last several years, decreasing about 5 percent between 2016 and 2018. The number of providers located in rural areas is not necessarily an indicator of access to care. The share of rural decedents using hospice has been increasing since 2000 (see Chart 11-9).
A Data Book: Health care spending and the Medicare program, July 2020 191
Chart 11-11. Hospice cases and length of stay, by diagnosis, 2018
Share of cases with Share length of stay Diagnosis of total cases greater than 180 days
Cancer 26% 9%
Alzheimer’s, nervous system disorders,
organic psychosis 23 35
Circulatory, except heart failure 20 25
Heart failure 8 23
Other 8 16
Respiratory disease 6 15
Chronic airway obstruction, NOS 5 29
Genitourinary disease 2 9
Digestive disease 2 9
All 100 21
Note: NOS (not otherwise specified). Cases include all patients who received hospice care in 2018, not just decedents. “Diagnosis” reflects primary diagnosis on the beneficiary’s last hospice claim in 2018. The share of cases with length of stay greater than 180 days reflects the share of hospice patients who received hospice care in 2018 whose lifetime length
of hospice stay exceeded 180 days at the end of 2018 (or at the time of death or discharge in 2018 if the beneficiary was not enrolled in hospice at the end of 2018).
Source: MedPAC analysis of 100 percent hospice claims standard analytic file from CMS and the Medicare Beneficiary Database.
• In 2018, the most common primary diagnoses among Medicare hospice patients were cancer (26 percent), neurological conditions (Alzheimer’s disease, nervous system disorders, and organic psychosis) (23 percent of cases), circulatory conditions other than heart failure (20 percent), and heart failure (8 percent).
• Length of stay varies by diagnosis. Long hospice stays were most common among patients with Alzheimer’s disease and other nervous system disorders, circulatory conditions (including heart failure), and chronic airway obstruction. Long hospice stays were least common among beneficiaries with cancer, genitourinary disease, and digestive disease.
192 Other services
Chart 11-12. Hospice average length of stay among decedents increased slightly in 2018
Average length of stay Percentiles of length of stay (in days)
Year (in days) 10th 25th 50th 75th 90th
2000 53.5 3 6 17 56 141
2016 87.0 2 5 17 78 243
2017 88.1 2 5 17 78 248
2018 89.6 2 5 18 81 253
Note: Length of stay is calculated for decedents who were using hospice at the time of death or before death and reflects the
total number of days the decedent was enrolled in the Medicare hospice benefit during his or her lifetime. Length-of-stay
data for 2016, 2017, and 2018 are based on the Medicare Beneficiary Database obtained from CMS in October 2019. Some length-of-stay figures for 2016 and 2017 differ from those published in the June 2019 data book because they were based on an earlier version of the Medicare Beneficiary Database obtained from CMS.
Source: MedPAC analysis of the denominator file and the Medicare Beneficiary Database from CMS.
• Average length of stay among decedents was 89.6 days in 2018, an increase from 2017 of more than one day.
• There is wide variation in hospice length of stay. In 2018, hospice length of stay among decedents ranged from 2 days at the 10th percentile to 253 days at the 90th percentile.
• Since 2000, growth in average length of stay among decedents has largely been the result of increases in length of stay for patients with the longest stays. Length of stay at the 90th percentile was more than 100 days greater in 2018 than in 2000.
• Short stays in hospice have changed little since 2000. For example, among decedents, median length of stay was 18 days in 2018 and 17 days in 2000. Hospice length of stay at the 25th percentile was 5 days in 2018 and 6 days in 2000.
A Data Book: Health care spending and the Medicare program, July 2020 193
Chart 11-13. Hospice length of stay among decedents, by beneficiary and hospice characteristics, 2018
Average length Length-of-stay percentiles (in days)
of stay (in days) 10th 50th 90th
Beneficiary Diagnosis Cancer 53 3 17 128 Neurological 151 4 38 445 Heart/circulatory 97 2 17 288 COPD 119 2 28 350 Other 56 2 8 156 Site of service Home 93 4 26 245 Nursing facility 106 3 21 310 Assisted living facility 155 5 54 438
Hospice For profit 110 3 23 321 Nonprofit 68 2 13 186 Freestanding 92 2 18 263 Home health based 70 2 15 191 Hospital based 57 2 12 153 Note: COPD (chronic obstructive pulmonary disease). Average length of stay is calculated for Medicare beneficiaries who died
in 2018 and used hospice that year, and it reflects the total number of days the decedent was enrolled in the Medicare hospice benefit during his or her lifetime. “Diagnosis” reflects primary diagnosis on the beneficiary’s last hospice claim.
Source: MedPAC analysis of 100 percent hospice claims standard analytic file data, Medicare Beneficiary Database, Medicare hospice cost reports, and Provider of Services file data from CMS.
• Hospice average length of stay among decedents varies by both beneficiary and provider characteristics. Most of this variation reflects differences in length of stay among patients with the longest stays (i.e., at the 90th percentile). Length of stay varies much less for patients with shorter stays (i.e., at the 10th or 50th percentile).
• Beneficiaries with neurological conditions and COPD have the longest stays, while beneficiaries with cancer have the shortest stays, on average.
• Beneficiaries who receive hospice services in assisted living facilities have longer stays on average than beneficiaries who receive care at home or in a nursing facility.
• For-profit and freestanding hospices have longer average lengths of stay than nonprofit and
provider-based (home health−based and hospital-based) hospices.
194 Other services
Chart 11-14. More than half of Medicare hospice spending in 2018 was for patients with stays exceeding 180 days Medicare hospice spending, 2018 (in billions) All hospice users in 2018 $19.2 Beneficiaries with LOS > 180 days 11.1 Days 1–180 3.8 Days 181–365 3.5 Days 366+ 3.8 Beneficiaries with LOS ≤ 180 days 8.2
Note: LOS (length of stay). LOS reflects the beneficiary’s lifetime LOS as of the end of 2018 (or at the time of death or discharge
in 2018 if the beneficiary was not enrolled in hospice at the end of 2018). All spending reflected in the chart occurred only in 2018. Break-out groups do not sum to total because of rounding.
Source: MedPAC analysis of 100 percent hospice claims standard analytic file data and the common Medicare enrollment file from CMS.
• In 2018, Medicare hospice spending on patients with stays exceeding 180 days was about $11.1 billion, more than half (58 percent) of all Medicare hospice spending that year.
• About $3.8 billion, or about 20 percent, of Medicare hospice spending in 2018 was on hospice care for patients who had already received at least one year of hospice.
A Data Book: Health care spending and the Medicare program, July 2020 195
Chart 11-15. Hospice aggregate Medicare margins, 2013–2017
Share of Medicare margin hospices (2017) 2013 2014 2015 2016 2017 All 100% 8.5% 8.2% 9.9% 10.9% 12.6% Freestanding 78 12.0 11.6 13.8 14.0 15.3 Home health based 11 2.5 3.5 3.3 6.2 8.0 Hospital based 10 –17.4 –20.8 –23.8 –16.7 –13.8 For profit 69 15.0 15.3 17.8 17.9 20.2 Nonprofit 27 0.8 –0.4 0.0 2.2 2.5 Government 4 N/A N/A N/A N/A N/A Urban 80 8.8 8.7 10.4 11.4 12.9 Rural 20 5.9 3.3 4.8 6.3 8.8 Below cap 86.0 8.6 8.4 9.9 10.7 12.5 Above cap 14.0 7.0 6.0 9.8 12.6 13.0 Above cap (including cap overpayments) 14.0 20.1 18.8 21.4 20.2 21.2
Note: N/A (not available). Margins for all provider categories exclude overpayments to above-cap hospices except where
specifically indicated. Margins are calculated based on Medicare-allowable, reimbursable costs. The percentages of
freestanding and provider-based (home health–based and hospital-based) hospices do not sum to 100 percent because skilled nursing facility–based hospices are not broken out separately.
Source: MedPAC analysis of Medicare hospice cost reports, 100 percent hospice claims standard analytic file, and Medicare Provider of Services data from CMS.
• The aggregate Medicare margin was 12.6 percent in 2017, up from 10.9 percent in 2016.
• In 2017, freestanding hospices had higher margins (15.3 percent) than home health–based (8.0 percent) and hospital-based hospices (–13.8 percent).
• The 2017 margin among for-profit hospices was high at 20.2 percent. Nonprofit hospices as a group had a margin of 2.5 percent in 2017, but the subset of nonprofit hospices that were freestanding had a higher margin, 5.7 percent (latter figure not shown in chart).
• The aggregate 2017 margin was higher for urban hospices (12.9 percent) than rural hospices (8.8 percent).
• Hospices that exceeded the cap (Medicare’s aggregate average per beneficiary payment limit) had a 2017 margin of about 21 percent before the return of the cap overpayments.
196 Other services
Chart 11-16. Medicare margins were higher among hospices with more long stays, 2017
Note: Margins exclude overpayments to hospices that exceeded the cap on the average annual Medicare payment per
beneficiary. Margins are calculated based on Medicare-allowable, reimbursable costs. For hospice providers in the lowest (first) quintile, the share of stays greater than 180 days was less than 12.4 percent; it was between 12.4 percent and 20.0 percent in the second quintile; it was between 20.0 percent and 26.7 percent in the third quintile; it was between 26.7
percent and 34.9 percent in the fourth quintile; and it was greater than 34.9 percent in the highest (fifth) quintile. Source: MedPAC analysis of Medicare hospice cost reports and 100 percent hospice claims standard analytic file from CMS.
• Medicare’s per diem payment system for hospice has provided an incentive for longer lengths of stay.
• Hospices with more patients who had stays greater than 180 days generally had higher margins in 2017. Hospices in the lowest length-of-stay quintile had a margin of –4.5 percent compared with a 22.1 percent margin for hospices in the second highest length-of-stay quintile.
(Chart continued next page)
-4.5
7.0
17.1
22.1
17.8
-10
-5
0
5
10
15
20
25
1 2 3 4 5
Marg
in (
in p
erc
en
t)
Share of providers' patients with stays greater than 180 days, by quintiles
A Data Book: Health care spending and the Medicare program, July 2020 197
Chart 11-16. Medicare margins were higher among hospices with more long stays, 2017 (continued)
• Margins were somewhat lower in the highest length-of-stay quintile (17.8 percent) compared with the second highest quintile (22.1 percent) because some hospices in the highest quintile exceeded Medicare’s aggregate payment cap and were required to repay the overage. Hospices exceeding the cap had a margin of about 20 percent before the return of overpayments (see Chart 11-15).
• The 2017 margin estimates reflect hospices’ financial performance in the second year of the new payment system, which began January 2016. The payment reforms modestly reduced the variation in profitability by length of stay. In 2015, there was a 29 percentage point spread in the margins between the lowest length of stay quintile (–8.9 percent) and the second highest length of stay quintile (20.4 percent) (data not shown). In 2017, the difference in margins between those length of stay quintiles narrowed to about 22 percentage points, as shown in the chart.
198 Other services
Chart 11-17. Hospices that exceeded Medicare’s annual payment cap, selected years
2002 2014 2015 2016 2017 Share of hospices exceeding the cap 2.6% 12.1% 12.3% 12.7% 14.0% Average payments over the cap per hospice exceeding the cap (in thousands) $470 $370 $316 $295 $273 Payments over the cap as a share of overall Medicare hospice spending 0.6% 1.2% 1.0% 1.0% 1.0%
Note: The aggregate cap statistics reflect the Commission’s estimates and may differ from those of the CMS claims-processing
contractors. Our estimates for 2014 to 2017 assume all hospices use the proportional methodology and rely on claims data through 14 months after the end of each cap year (with the exception of 2017, which used 15 months). The claims-processing contractors may reopen the hospice cap calculation for up to three years; the reopening process and timing
may vary across contractors. To illustrate the potential effect of reopening, we re-estimated cap overpayments for 2014 and 2015 using 38 months of claims data after the end of each cap year. With 38 months of data, the estimated share of hospices exceeding the cap increased by roughly 1 percentage point, and the average payments over the cap per
hospice exceeding the cap increased by roughly $20,000 in both 2014 and 2015. Cap year 2017 reflects an 11-month period from November 1, 2016, to Sept 30, 2017. For years before 2017, the cap year was defined as the period beginning November 1 and ending October 31 of the following year.
Source: MedPAC analysis of 100 percent hospice claims standard analytic file data, Medicare hospice cost reports, Provider of
Services file data from CMS, and CMS Providing Data Quickly system. Data on total spending for each fiscal year are
from the CMS Office of the Actuary or MedPAC estimates.
• The share of hospices exceeding the aggregate cap was 14.0 percent in 2017, up from 12.7 percent in 2016.
• Medicare payments over the cap represented 1.0 percent of total Medicare hospice spending in 2017.
• On average, above-cap hospices exceeded the cap by about $273,000 per provider in 2017, down from about $295,000 per provider in 2016.
A Data Book: Health care spending and the Medicare program, July 2020 199
Chart 11-18. Hospice live-discharge rates, 2016–2018 2016 2017 2018
Live discharge as a share of all discharges, by reason for live discharge All live discharges 16.9% 16.7% 17.0%
No longer terminally ill 6.8 6.5 6.3 Beneficiary revocation 6.4 6.4 6.6 Transfer hospice providers 2.1 2.1 2.2 Move out of service area 1.2 1.4 1.6
Discharge for cause 0.3 0.3 0.3
Providers’ overall rate of live discharge as a share of all discharges, by percentile (for providers with more than 30 discharges) 10th percentile 8.6 8.5 8.5 25th percentile 11.8 12.2 12.0 50th percentile 17.6 18.1 17.9 75th percentile 26.7 27.1 27.8 90th percentile 40.8 41.4 42.5
Note: Percentages may not sum to totals due to rounding. “All discharges” includes patients discharged alive or deceased. Source: MedPAC analysis of 100 percent hospice claims standard analytic file.
• In 2018, the overall live-discharge rate was 17.0 percent and has changed little since 2016.
• The most common reasons for live discharge were the beneficiary revoking the hospice benefit and the beneficiary no longer being terminally ill, accounting for 39 percent and 37 percent of live discharges, respectively. Less frequent reasons for live discharges included a beneficiary transferring hospice providers, a beneficiary moving out of the service area, and a beneficiary being discharged for cause.
• Among providers with more than 30 discharges, 10 percent of providers had live-discharge rates in excess of 42 percent.
• Small hospices as a group have substantially higher live-discharge rates than larger hospices. In 2018, the aggregate live-discharge rate was 44 percent for hospices with 30 or fewer discharges (data not shown).
200 Other services
Chart 11-19. Medicare spending for clinical laboratory services, 2005–2019
Note: Spending is for services paid under the clinical laboratory fee schedule. Hospital-based services are furnished in labs owned or operated by hospitals. The components of each bar may not sum to the total at the top of each bar due to
rounding. The spending data include only program payments; there is no beneficiary cost sharing for clinical lab services.
Source: The annual report of the Boards of Trustees of the Medicare trust funds, 2015 and 2020.
• Medicare spending for clinical laboratory services in all settings grew by an average of 3.6 percent per year between 2005 and 2013.
• From 2013 to 2014, Medicare spending for lab services declined by about 9 percent because, beginning in 2014, many lab tests provided in hospital outpatient departments are no longer paid separately under the clinical lab fee schedule. Instead, many of these tests are packaged with their associated visits or procedures under the hospital outpatient prospective payment system.
• Medicare spending for lab services decreased by an average of 0.9 percent per year from 2014 to 2017.
• Beginning in 2018, clinical laboratory fee schedule payment rates are based on private sector rates. From 2017 to 2019, Medicare spending for lab services grew by an average of 3.8 percent per year.
3.6 3.7 4.0 4.34.7 4.9 4.7
5.2 5.35.7 5.7
5.2 5.35.8 6.1
3.43.5
3.63.7
3.53.6 3.8
4.0 4.02.7 2.8
2.9 2.9
2.92.8
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
11.0
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Do
llars
(in
bil
lio
ns)
Calendar year
Independent and physician office Hospital based
8.58.1
8.78.2
8.9
7.3 7.6
7.0
8.0 8.2
8.6 8.5
9.3 9.3
8.5